NEsper - Event Stream and Complex Event Processing for .NET

Reference Documentation

4.1.0


Table of Contents

Preface
1. Technology Overview
1.1. Introduction to CEP and event stream analysis
1.2. CEP and relational databases
1.3. The Esper engine for CEP
1.4. Required 3rd Party Libraries
2. Event Representations
2.1. Event Underlying Objects
2.2. Event Properties
2.2.1. Escape Characters
2.3. Dynamic Event Properties
2.4. Fragment and Fragment Type
2.5. Object Events
2.5.1. Object Event Properties
2.5.2. Property Names
2.5.3. Constants and Enumeration
2.5.4. Parameterized Types
2.5.5. Known Limitations
2.6. Map Events
2.6.1. Overview
2.6.2. Map Properties
2.6.3. Map Supertypes
2.6.4. Advanced Map Property Types
2.6.4.1. Nested Properties
2.6.4.2. Map Event Type Properties
2.6.4.3. One-to-Many Relationships
2.7. System.Xml.XmlNode XML Events
2.7.1. Schema-Provided XML Events
2.7.1.1. Getting Started
2.7.1.2. Property Expressions and Namespaces
2.7.1.3. Property Expression to XPath Rewrite
2.7.1.4. Array Properties
2.7.1.5. Dynamic Properties
2.7.1.6. Transposing Properties
2.7.1.7. Event Sender
2.7.2. No-Schema-Provided XML Events
2.7.3. Explicitly-Configured Properties
2.7.3.1. Simple Explicit Property
2.7.3.2. Explicit Property Casting and Parsing
2.7.3.3. Node and Nodeset Explicit Property
2.8. Additional Event Representations
2.9. Updating, Merging and Versioning Events
2.10. Coarse-Grained Events
2.11. Event Objects Populated by Insert Into
3. Processing Model
3.1. Introduction
3.2. Insert Stream
3.3. Insert and Remove Stream
3.4. Filters and Where-clauses
3.5. Time Windows
3.5.1. Time Window
3.5.2. Time Batch
3.6. Batch Windows
3.7. Aggregation and Grouping
3.7.1. Insert and Remove Stream
3.7.2. Output for Aggregation and Group-By
3.7.2.1. Un-aggregated and Un-grouped
3.7.2.2. Fully Aggregated and Un-grouped
3.7.2.3. Aggregated and Un-Grouped
3.7.2.4. Fully Aggregated and Grouped
3.7.2.5. Aggregated and Grouped
3.8. Event Visibility and Current Time
4. EPL Reference: Clauses
4.1. EPL Introduction
4.2. EPL Syntax
4.2.1. Specifying Time Periods
4.2.2. Using Comments
4.2.3. Reserved Keywords
4.2.4. Escaping Strings
4.2.5. Data Types
4.2.5.1. Data Type of Constants
4.2.6. Annotation
4.2.6.1. Application-Provided Annotations
4.2.6.2. Built-In Annotations
4.2.6.3. @Name
4.2.6.4. @Description
4.2.6.5. @Tag
4.2.6.6. @Priority
4.2.6.7. @Drop
4.2.6.8. @Hint
4.2.6.9. @Hook
4.3. Choosing Event Properties And Events: the Select Clause
4.3.1. Choosing all event properties: select *
4.3.2. Choosing specific event properties
4.3.3. Expressions
4.3.4. Renaming event properties
4.3.5. Choosing event properties and events in a join
4.3.6. Choosing event properties and events from a pattern
4.3.7. Selecting insert and remove stream events
4.3.8. Qualifying property names and stream names
4.3.9. Select Distinct
4.4. Specifying Event Streams: the From Clause
4.4.1. Filter-based Event Streams
4.4.1.1. Specifying an Event Type
4.4.1.2. Specifying Filter Criteria
4.4.1.3. Filtering Ranges
4.4.1.4. Filtering Sets of Values
4.4.1.5. Filter Limitations
4.4.2. Pattern-based Event Streams
4.4.3. Specifying Views
4.4.4. Multiple Data Window Views
4.4.5. Using the Stream Name
4.5. Specifying Search Conditions: the Where Clause
4.6. Aggregates and grouping: the Group-by Clause and the Having Clause
4.6.1. Using aggregate functions
4.6.2. Organizing statement results into groups: the Group-by clause
4.6.3. Selecting groups of events: the Having clause
4.6.4. How the stream filter, Where, Group By and Having clauses interact
4.6.5. Comparing the Group By clause and the std:groupwin view
4.7. Stabilizing and Controlling Output: the Output Clause
4.7.1. Output Clause Options
4.7.1.1. Controlling Output Using an Expression
4.7.1.2. Suppressing Output With After
4.7.2. Aggregation, Group By, Having and Output clause interaction
4.7.3. Runtime Considerations
4.8. Sorting Output: the Order By Clause
4.9. Limiting Row Count: the Limit Clause
4.10. Merging Streams and Continuous Insertion: the Insert Into Clause
4.10.1. Transposing a Property To a Stream
4.10.2. Merging Streams By Event Type
4.10.3. Merging Disparate Types of Events: Variant Streams
4.10.4. Decorated Events
4.10.5. Event as a Property
4.10.6. Populating an Underlying Event Object
4.11. Joining Event Streams
4.12. Outer and Inner Joins
4.13. Unidirectional Joins
4.14. Subqueries
4.14.1. The 'exists' Keyword
4.14.2. The 'in' and 'not in' Keywords
4.14.3. The 'any' and 'some' Keywords
4.14.4. The 'all' Keyword
4.14.5. Multi-Column Selection
4.14.6. Multi-Row Selection
4.14.7. Hints Related to Subqueries
4.15. Accessing Relational Data via SQL
4.15.1. Joining SQL Query Results
4.15.2. SQL Query and the EPL Where Clause
4.15.3. Outer Joins With SQL Queries
4.15.4. Using Patterns to Request (Poll) Data
4.15.5. Polling SQL Queries via Iterator
4.15.6. ADO.NET Implementation Overview
4.15.7. No-Metadata Workaround
4.15.8. SQL Input Parameter and Column Output Conversion
4.15.9. SQL Row Object Conversion
4.16. Accessing Non-Relational Data via Method Invocation
4.16.1. Joining Method Invocation Results
4.16.2. Polling Method Invocation Results via Iterator
4.16.3. Providing the Method
4.16.4. Using a Map Return Type
4.17. Creating and Using Named Windows
4.17.1. Creating Named Windows: the Create Window clause
4.17.1.1. Creation by Modelling after an Existing Type
4.17.1.2. Creation By Defining Columns Names and Types
4.17.1.3. Dropping or Removing Named Windows
4.17.2. Inserting Into Named Windows
4.17.2.1. Named Windows Holding Decorated Events
4.17.2.2. Named Windows Holding Events As Property
4.17.3. Selecting From Named Windows
4.17.4. Triggered Select on Named Windows: the On Select clause
4.17.5. Triggered Playback from Named Windows: the On Insert clause
4.17.6. Populating a Named Window from an Existing Named Window
4.17.7. Updating Named Windows: the On Update clause
4.17.8. Deleting From Named Windows: the On Delete clause
4.17.8.1. Using Patterns in the On Delete Clause
4.17.9. Triggered Upsert using the On-Merge Clause
4.17.10. Explicitly Indexing Named Windows
4.17.11. Versioning and Revision Event Type Use with Named Windows
4.18. Declaring an Event Type: Create Schema
4.18.1. Declare an Event Type by Providing Names and Types
4.18.2. Declare an Event Type by Providing a Class Name
4.18.3. Declare a Variant Stream
4.19. Splitting and Duplicating Streams
4.20. Variables
4.20.1. Creating Variables: the Create Variable clause
4.20.2. Setting Variable Values: the On Set clause
4.20.3. Using Variables
4.20.4. Object-Type Variables
4.20.5. Class and Event-Type Variables
4.21. Contained-Event Selection
4.21.1. Select Clause in a Contained-Event Selection
4.21.2. Where Clause in a Contained-Event Selection
4.21.3. Contained-Event Selection and Joins
4.22. Updating an Insert Stream: the Update IStream Clause
4.22.1. Immutability and Updates
4.23. Controlling Event Delivery : The For Clause
5. EPL Reference: Patterns
5.1. Event Pattern Overview
5.2. How to use Patterns
5.2.1. Pattern Syntax
5.2.2. Patterns in EPL
5.2.3. Subscribing to Pattern Events
5.2.4. Pulling Data from Patterns
5.2.5. Pattern Error Reporting
5.3. Operator Precedence
5.4. Filter Expressions In Patterns
5.5. Pattern Operators
5.5.1. Every
5.5.1.1. Limiting Subexpression Lifetime
5.5.1.2. Every Operator Example
5.5.1.3. Sensor Example
5.5.2. Every-Distinct
5.5.3. Repeat
5.5.4. Repeat-Until
5.5.4.1. Unbound Repeat
5.5.4.2. Bound Repeat Overview
5.5.4.3. Bound Repeat - Open Ended Range
5.5.4.4. Bound Repeat - High Endpoint Range
5.5.4.5. Bound Repeat - Bounded Range
5.5.4.6. Tags and the Repeat Operator
5.5.5. And
5.5.6. Or
5.5.7. Not
5.5.8. Followed-by
5.5.8.1. Limiting Sub-Expression Count
5.5.9. Pattern Guards
5.5.9.1. The timer:within Pattern Guard
5.5.9.2. The timer:withinmax Pattern Guard
5.5.9.3. The while Pattern Guard
5.5.9.4. Guard Time Interval Expressions
5.5.9.5. Combining Guard Expressions
5.6. Pattern Atoms
5.6.1. Filter Atoms
5.6.2. Time-based Observer Atoms
5.6.2.1. timer:interval
5.6.2.2. timer:at
6. EPL Reference: Match Recognize
6.1. Overview
6.2. Comparison of Match Recognize and EPL Patterns
6.3. Syntax
6.3.1. Syntax Example
6.4. Pattern and Pattern Operators
6.4.1. Operator Precedence
6.4.2. Concatenation
6.4.3. Alternation
6.4.4. Quantifiers Overview
6.4.5. Variables Can be Singleton or Group
6.4.5.1. Additional Aggregation Functions
6.4.6. Eliminating Duplicate Matches
6.4.7. Greedy Or Reluctant
6.4.8. Quantifier - One Or More (+ and +?)
6.4.9. Quantifier - Zero Or More (* and *?)
6.4.10. Quantifier - Zero Or One (? and ??)
6.5. Define Clause
6.5.1. The Prev Operator
6.6. Measure Clause
6.7. Datawindow-Bound
6.8. Interval
6.9. Limitations
7. EPL Reference: Operators
7.1. Arithmetic Operators
7.2. Logical And Comparison Operators
7.3. Concatenation Operators
7.4. Binary Operators
7.5. Array Definition Operator
7.6. Dot Operator
7.6.1. Duck Typing
7.7. The 'in' Keyword
7.8. The 'between' Keyword
7.9. The 'like' Keyword
7.10. The 'regexp' Keyword
7.11. The 'any' and 'some' Keywords
7.12. The 'all' Keyword
8. EPL Reference: Functions
8.1. Single-row Function Reference
8.1.1. The Case Control Flow Function
8.1.2. The Cast Function
8.1.3. The Coalesce Function
8.1.4. The Current_Timestamp Function
8.1.5. The Exists Function
8.1.6. The Instance-Of Function
8.1.7. The Min and Max Functions
8.1.8. The Previous Function
8.1.8.1. Restrictions
8.1.8.2. Comparison to the prior Function
8.1.9. The Previous-Tail Function
8.1.9.1. Restrictions
8.1.10. The Previous-Window Function
8.1.10.1. Restrictions
8.1.11. The Previous-Count Function
8.1.11.1. Restrictions
8.1.12. The Prior Function
8.1.13. The Type-Of Function
8.2. Aggregate Functions
8.2.1. SQL-Standard Functions
8.2.2. Data Window Aggregation Functions
8.2.2.1. First Aggregation Function
8.2.2.2. Last Aggregation Function
8.2.2.3. Window Aggregation Function
8.2.3. Additional Aggregation Functions
8.3. User-Defined Functions
9. EPL Reference: Views
9.1. Window views
9.1.1. Length window (win:length)
9.1.2. Length batch window (win:length_batch)
9.1.3. Time window (win:time)
9.1.4. Externally-timed window (win:ext_timed)
9.1.5. Time batch window (win:time_batch)
9.1.6. Time-Length combination batch window (win:time_length_batch)
9.1.7. Time-Accumulating window (win:time_accum)
9.1.8. Keep-All window (win:keepall)
9.1.9. First Length (win:firstlength)
9.1.10. First Time (win:firsttime)
9.2. Standard view set
9.2.1. Unique (std:unique)
9.2.2. Grouped Data Window (std:groupwin)
9.2.3. Size (std:size)
9.2.4. Last Event (std:lastevent)
9.2.5. First Event (std:firstevent)
9.2.6. First Unique (std:firstunique)
9.3. Statistics views
9.3.1. Univariate statistics (stat:uni)
9.3.2. Regression (stat:linest)
9.3.3. Correlation (stat:correl)
9.3.4. Weighted average (stat:weighted_avg)
9.4. Extension View Set
9.4.1. Sorted Window View (ext:sort)
9.4.2. Time-Order View (ext:time_order)
10. API Reference
10.1. API Overview
10.2. The Service Provider Interface
10.3. The Administrative Interface
10.3.1. Creating Statements
10.3.2. Receiving Statement Results
10.3.3. Setting a Subscriber Object
10.3.3.1. Row-By-Row Delivery
10.3.3.2. Multi-Row Delivery
10.3.4. Adding Event Handlers
10.3.4.1. Subscription Snapshot and Atomic Delivery
10.3.5. Using Enumerators
10.3.6. Managing Statements
10.3.7. Runtime Configuration
10.4. The Runtime Interface
10.4.1. Event Sender
10.4.2. Receiving Unmatched Events
10.4.3. On-Demand Snapshot Query Execution
10.4.3.1. On-Demand Query API
10.5. Event and Event Type
10.5.1. Event Type Metadata
10.5.2. Event Object
10.5.3. Query Example
10.5.4. Pattern Example
10.6. Engine Threading and Concurrency
10.6.1. Advanced Threading
10.6.1.1. Inbound Threading
10.6.1.2. Outbound Threading
10.6.1.3. Timer Execution Threading
10.6.1.4. Route Execution Threading
10.6.1.5. Threading Service Provider Interface
10.6.2. Processing Order
10.6.2.1. Competing Statements
10.6.2.2. Competing Events in a Work Queue
10.7. Controlling Time-Keeping
10.7.1. Controlling Time Using Time Span Events
10.7.2. Additional Time-Related APIs
10.8. Time Resolution
10.9. Service Isolation
10.9.1. Overview
10.9.2. Example: Suspending a Statement
10.9.3. Example: Catching up a Statement from Historical Data
10.9.4. Isolation for Insert-Into
10.9.5. Isolation for Named Windows
10.9.6. Runtime Considerations
10.10. Exception Handling
10.11. Condition Handling
10.12. Statement Object Model
10.12.1. Building an Object Model
10.12.2. Building Expressions
10.12.3. Building a Pattern Statement
10.12.4. Building a Select Statement
10.12.5. Building a Create-Variable and On-Set Statement
10.12.6. Building Create-Window, On-Delete and On-Select Statements
10.13. Prepared Statement and Substitution Parameters
10.14. Engine and Statement Metrics Reporting
10.14.1. Engine Metrics
10.14.2. Statement Metrics
10.15. Event Rendering to XML and JSON
10.15.1. JSON Event Rendering Conventions and Options
10.15.2. XML Event Rendering Conventions and Options
10.16. Plug-in Loader
10.17. CLR-Java Differences
11. Configuration
11.1. Programmatic Configuration
11.2. Configuration via XML File
11.3. XML Configuration File
11.4. Configuration Items
11.4.1. Events represented by CLR types
11.4.1.1. Package of CLR Event Classes
11.4.1.2. Event type name to CLR type mapping
11.4.1.3. Legacy Event Classes
11.4.1.4. Specifying Event Properties for CLR Types
11.4.1.5. Turning off Code Generation
11.4.1.6. Case Sensitivity and Property Names
11.4.1.7. Factory and Copy Method
11.4.2. Events represented by System.Collection.Generic.IDictionary
11.4.3. Events represented by System.Xml.XmlNode
11.4.3.1. Schema Resource
11.4.3.2. Explicit XPath Property
11.4.3.3. Absolute or Deep Property Resolution
11.4.3.4. XPath Variable and Function Resolver
11.4.3.5. Auto Fragment
11.4.3.6. XPath Property Expression
11.4.3.7. Event Sender Setting
11.4.4. Events represented by Plug-in Event Representations
11.4.4.1. Enabling an Custom Event Representation
11.4.4.2. Adding Plug-in Event Types
11.4.4.3. Setting Resolution URIs
11.4.5. Class and package imports
11.4.6. Cache Settings for From-Clause Method Invocations
11.4.7. Variables
11.4.8. Relational Database Access
11.4.8.1. Connections obtained via DataSource
11.4.8.2. Connections obtained via DataSource Factory
11.4.8.3. Connections obtained via DriverManager
11.4.8.4. Connections-level settings
11.4.8.5. Connections lifecycle settings
11.4.8.6. Cache settings
11.4.8.7. Column Change Case
11.4.8.8. SQL Types Mapping
11.4.8.9. Metadata Origin
11.4.9. Engine Settings related to Concurrency and Threading
11.4.9.1. Preserving the order of events delivered to listeners
11.4.9.2. Preserving the order of events for insert-into streams
11.4.9.3. Internal Timer Settings
11.4.9.4. Advanced Threading Options
11.4.10. Engine Settings related to Event Metadata
11.4.10.1. Type Property Names, Case Sensitivity and Accessor Style
11.4.11. Engine Settings related to View Resources
11.4.11.1. Sharing View Resources between Statements
11.4.11.2. Configuring Multi-Expiry Policy Defaults
11.4.12. Engine Settings related to Logging
11.4.12.1. Execution Path Debug Logging
11.4.12.2. Query Plan Logging
11.4.12.3. JDBC Logging
11.4.13. Engine Settings related to Variables
11.4.13.1. Variable Version Release Interval
11.4.14. Engine Settings related to Stream Selection
11.4.14.1. Default Statement Stream Selection
11.4.15. Engine Settings related to Time Source
11.4.15.1. Default Time Source
11.4.16. Engine Settings related to Metrics Reporting
11.4.17. Engine Settings related to Language and Locale
11.4.18. Engine Settings related to Expression Evaluation
11.4.18.1. Integer Division and Division by Zero
11.4.18.2. Subselect Evaluation Order
11.4.18.3. User-Defined Function or Static Method Cache
11.4.18.4. Extended Built-in Aggregation Functions
11.4.18.5. Duck Typing
11.4.19. Engine Settings related to Execution of Statements
11.4.19.1. Prioritized Execution
11.4.20. Engine Settings related to Exception Handling
11.4.21. Engine Settings related to Condition Handling
11.4.22. Revision Event Type
11.4.23. Variant Stream
11.5. Type Names
11.6. Runtime Configuration
11.7. Logging Configuration
11.7.1. Log4net Logging Configuration
12. Packaging and Deploying
12.1. Overview
12.2. EPL Modules
12.3. The Deployment Administrative Interface
12.3.1. Reading Module Content
12.3.2. Ordering Multiple Modules
12.3.3. Deploying and Undeploying
12.3.4. Listing Deployments
12.3.5. State Transitioning a Module
12.3.6. Best Practices
13. Extension and Plug-in
13.1. Custom Single-Row Functions
13.1.1. Implementing a Single-Row Function
13.1.2. Configuring the Single-Row Function Name
13.2. Custom View Implementation
13.2.1. Implementing a View Factory
13.2.2. Implementing a View
13.2.3. View Contract
13.2.4. Configuring View Namespace and Name
13.2.5. Requirement for Data Window Views
13.2.6. Requirement for Grouped Views
13.3. Custom Aggregation Functions
13.3.1. Implementing an Aggregation Function
13.3.2. Configuring the Aggregation Function Name
13.3.3. Accepting Multiple Parameters
13.4. Custom Pattern Guard
13.4.1. Implementing a Guard Factory
13.4.2. Implementing a Guard Class
13.4.3. Configuring Guard Namespace and Name
13.5. Custom Pattern Observer
13.5.1. Implementing an Observer Factory
13.5.2. Implementing an Observer Class
13.5.3. Configuring Observer Namespace and Name
13.6. Custom Event Representation
13.6.1. How It Works
13.6.2. Steps
13.6.3. URI-based Resolution
13.6.4. Example
13.6.4.1. Sample Event Type
13.6.4.2. Sample Event Bean
13.6.4.3. Sample Event Representation
13.6.4.4. Sample Event Bean Factory
14. Examples, Tutorials, Case Studies
14.1. Examples Overview
14.2. Running the Examples
14.3. AutoID RFID Reader
14.4. Runtime Configuration
14.5. Market Data Feed Monitor
14.5.1. Input Events
14.5.2. Computing Rates Per Feed
14.5.3. Detecting a Fall-off
14.5.4. Event generator
14.6. OHLC Plug-in View
14.7. Transaction 3-Event Challenge
14.7.1. The Events
14.7.2. Combined event
14.7.3. Real time summary data
14.7.4. Find problems
14.7.5. Event generator
14.8. Self-Service Terminal
14.8.1. Events
14.8.2. Detecting Customer Check-in Issues
14.8.3. Absence of Status Events
14.8.4. Activity Summary Data
14.9. Assets Moving Across Zones - An RFID Example
14.10. StockTicker
14.11. MatchMaker
14.12. Named Window Query
14.13. Quality of Service
15. Performance
15.1. Performance Results
15.2. Performance Tips
15.2.1. Understand how to tune your application
15.2.2. Compare Esper to other solutions
15.2.3. Input and Output Bottlenecks
15.2.4. Advanced Theading
15.2.5. Select the underlying event rather than individual fields
15.2.6. Prefer stream-level filtering over post-data-window filtering
15.2.7. Reduce the use of arithmetic in expressions
15.2.8. Remove Unneccessary Constructs
15.2.9. End Pattern Sub-Expressions
15.2.10. Consider using EventPropertyGetter for fast access to event properties
15.2.11. Consider casting the underlying event
15.2.12. Turn off logging
15.2.13. Disable view sharing
15.2.14. Tune or disable delivery order guarantees
15.2.15. Use a Subscriber Object to Receive Events
15.2.16. High-Arrival-Rate Streams and Single Statements
15.2.17. Joins And Where-clause And Data Windows
15.2.18. Patterns and Pattern Sub-Expression Instances
15.2.19. The Keep-All Data Window
15.2.20. Statement Design for Reduced Memory Consumption
15.2.21. Performance, JVM, OS and hardware
15.2.22. Consider using Hints
15.2.23. Optimizing Stream Filter Expressions
15.3. Using the performance kit
15.3.1. How to use the performance kit
15.3.2. How we use the performance kit
16. References
16.1. Reference List
A. Output Reference and Samples
A.1. Introduction and Sample Data
A.2. Output for Un-aggregated and Un-grouped Queries
A.2.1. No Output Rate Limiting
A.2.2. Output Rate Limiting - Default
A.2.3. Output Rate Limiting - Last
A.2.4. Output Rate Limiting - First
A.2.5. Output Rate Limiting - Snapshot
A.3. Output for Fully-aggregated and Un-grouped Queries
A.3.1. No Output Rate Limiting
A.3.2. Output Rate Limiting - Default
A.3.3. Output Rate Limiting - Last
A.3.4. Output Rate Limiting - First
A.3.5. Output Rate Limiting - Snapshot
A.4. Output for Aggregated and Un-grouped Queries
A.4.1. No Output Rate Limiting
A.4.2. Output Rate Limiting - Default
A.4.3. Output Rate Limiting - Last
A.4.4. Output Rate Limiting - First
A.4.5. Output Rate Limiting - Snapshot
A.5. Output for Fully-aggregated and Grouped Queries
A.5.1. No Output Rate Limiting
A.5.2. Output Rate Limiting - Default
A.5.3. Output Rate Limiting - All
A.5.4. Output Rate Limiting - Last
A.5.5. Output Rate Limiting - First
A.5.6. Output Rate Limiting - Snapshot
A.6. Output for Aggregated and Grouped Queries
A.6.1. No Output Rate Limiting
A.6.2. Output Rate Limiting - Default
A.6.3. Output Rate Limiting - All
A.6.4. Output Rate Limiting - Last
A.6.5. Output Rate Limiting - First
A.6.6. Output Rate Limiting - Snapshot
B. Reserved Keywords
Index

Preface

Analyzing and reacting to information in real-time oftentimes requires the development of custom applications. Typically these applications must obtain the data to analyze, filter data, derive information and then indicate this information through some form of presentation or communication. Data may arrive with high frequency requiring high throughput processing. And applications may need to be flexible and react to changes in requirements while the data is processed. Esper is an event stream processor that aims to enable a short development cycle from inception to production for these types of applications.

This document is a resource for software developers who develop event driven applications. It also contains information that is useful for business analysts and system architects who are evaluating Esper.

It is assumed that the reader is familiar with the CLR and the programming languages that can be used to interoperate with the CLR.

This document is relevant in all phases of your software development project: from design to deployment and support.

If you are new to Esper, please follow these steps:

  1. Read the tutorials, case studies and solution patterns available on the Esper public web site at http://esper.codehaus.org

  2. Read Section 1.1, “Introduction to CEP and event stream analysis” if you are new to CEP and ESP (complex event processing, event stream processing)

  3. Read Chapter 2, Event Representations that explains the different ways of representing events to Esper

  4. Read Chapter 3, Processing Model to gain insight into EPL continuous query results

  5. Read Section 4.1, “EPL Introduction” for an introduction to event stream processing via EPL

  6. Read Section 5.1, “Event Pattern Overview” for an overview over event patterns

  7. Read Section 6.1, “Overview” for an overview over event patterns using the match recognize syntax.

  8. Then glance over the examples Section 14.1, “Examples Overview”

  9. Finally to test drive Esper performance, read Chapter 15, Performance

Chapter 1. Technology Overview

1.1. Introduction to CEP and event stream analysis

The Esper engine has been developed to address the requirements of applications that analyze and react to events. Some typical examples of applications are:

  • Business process management and automation (process monitoring, BAM, reporting exceptions)

  • Finance (algorithmic trading, fraud detection, risk management)

  • Network and application monitoring (intrusion detection, SLA monitoring)

  • Sensor network applications (RFID reading, scheduling and control of fabrication lines, air traffic)

What these applications have in common is the requirement to process events (or messages) in real-time or near real-time. This is sometimes referred to as complex event processing (CEP) and event stream analysis. Key considerations for these types of applications are throughput, latency and the complexity of the logic required.

  • High throughput - applications that process large volumes of messages (between 1,000 to 100k messages per second)

  • Low latency - applications that react in real-time to conditions that occur (from a few milliseconds to a few seconds)

  • Complex computations - applications that detect patterns among events (event correlation), filter events, aggregate time or length windows of events, join event streams, trigger based on absence of events etc.

The Esper engine was designed to make it easier to build and extend CEP applications.

1.2. CEP and relational databases

Relational databases and the standard query language (SQL) are designed for applications in which most data is fairly static and complex queries are less frequent. Also, most databases store all data on disks (except for in-memory databases) and are therefore optimized for disk access.

To retrieve data from a database an application must issue a query. If an application need the data 10 times per second it must fire the query 10 times per second. This does not scale well to hundreds or thousands of queries per second.

Database triggers can be used to fire in response to database update events. However database triggers tend to be slow and often cannot easily perform complex condition checking and implement logic to react.

In-memory databases may be better suited to CEP applications than traditional relational database as they generally have good query performance. Yet they are not optimized to provide immediate, real-time query results required for CEP and event stream analysis.

1.3. The Esper engine for CEP

The Esper engine works a bit like a database turned upside-down. Instead of storing the data and running queries against stored data, the Esper engine allows applications to store queries and run the data through. Response from the Esper engine is real-time when conditions occur that match queries. The execution model is thus continuous rather than only when a query is submitted.

Esper provides two principal methods or mechanisms to process events: event patterns and event stream queries.

Esper offers an event pattern language to specify expression-based event pattern matching. Underlying the pattern matching engine is a state machine implementation. This method of event processing matches expected sequences of presence or absence of events or combinations of events. It includes time-based correlation of events.

Esper also offers event stream queries that address the event stream analysis requirements of CEP applications. Event stream queries provide the windows, aggregation, joining and analysis functions for use with streams of events. These queries are following the EPL syntax. EPL has been designed for similarity with the SQL query language but differs from SQL in its use of views rather than tables. Views represent the different operations needed to structure data in an event stream and to derive data from an event stream.

Esper provides these two methods as alternatives through the same API.

1.4. Required 3rd Party Libraries

Esper requires the following 3rd-party libraries at runtime:

  • ANTLR is the parser generator used for parsing and parse tree walking of the pattern and EPL syntax. Credit goes to Terence Parr at http://www.antlr.org. The ANTLR license is in the lib directory. The library is required for compile-time only.

  • Log4net is a logging API. This open source software is under the Apache license. The Apache 2.0 license is in the lib directory.

Esper requires the following 3rd-party libraries at compile-time and for running the test suite:

  • NUnit is a unit testing framework designed for .NET. Its license has also been placed in the lib directory. The library is required for build-time only.

Chapter 2. Event Representations

This section outlines the different means to model and represent events.

Esper uses the term event type to describe the type information available for an event representation.

Your application may configure predefined event types at startup time or dynamically add event types at runtime via API or EPL syntax. See Section 11.4, “Configuration Items” for startup-time configuration and Section 10.3.7, “Runtime Configuration” for the runtime configuration API.

The EPL create schema syntax allows declaring an event type at runtime using EPL, see Section 4.18, “Declaring an Event Type: Create Schema”.

In Section 10.5, “Event and Event Type” we explain how an event type becomes visible in EPL statements and output events delivered by the engine.

2.1. Event Underlying Objects

An event is an immutable record of a past occurrence of an action or state change. Event properties capture the state information for an event.

In Esper, an event can be represented by any of the following underlying objects:

Table 2.1. Event Underlying Vanilla Objects

CLR TypeDescription
System.ObjectAny CLR type with read accessible properties.
System.Collection.Generic.IDictionary<string, object>Map (dictionary) events are key-values pairs and can also contain objects, further Map, and arrays thereof.
System.Xml.XmlNodeXML document object model (DOM).
Application classesPlug-in event representation via the extension API.

Esper provides multiple choices for representing an event. There is no absolute need for you to create new types to represent an event.

Event representations have the following in common:

  • All event representations support nested, indexed and mapped properties (aka. property expression), as explained in more detail below. There is no limitation to the nesting level.

  • All event representations provide event type metadata. This includes type metadata for nested properties.

  • All event representations allow transposing the event itself and parts of all of its property graph into new events. The term transposing refers to selecting the event itself or event properties that are themselves nestable property graphs, and then querying the event's properties or nested property graphs in further statements. The Apache Axiom event representation is an exception and does not currently allow transposing event properties but does allow transposing the event itself.

  • The object and Map representations allow supertypes.

The API behavior for all event representations is the same, with minor exceptions noted in this chapter.

The benefits of multiple event representations are:

  • For applications that already have events in one of the supported representations, there is no need to transform events into an object before processing.

  • Event representations are exchangeable, reducing or eliminating the need to change statements when the event representation changes.

  • Event representations are interoperable, allowing all event representations to interoperate in same or different statements.

  • The choice makes its possible to consciously trade-off performance, ease-of-use, the ability to evolve and effort needed to import or externalize events and use existing event type metadata.

2.2. Event Properties

Event properties capture the state information for an event. Event properties be simple as well as indexed, mapped and nested event properties. The table below outlines the different types of properties and their syntax in an event expression. This syntax allows statements to query deep objects graphs, XML structures and Map events.

Table 2.2. Types of Event Properties

TypeDescriptionSyntaxExample
SimpleA property that has a single value that may be retrieved.
name
sensorId
IndexedAn indexed property stores an ordered collection of objects (all of the same type) that can be individually accessed by an integer-valued, non-negative index (or subscript).
name[index]
sensor[0]
MappedA mapped property stores a keyed collection of objects (all of the same type).
name('key')
sensor('light')
NestedA nested property is a property that lives within another property of an event.
name.nestedname
sensor.value

Combinations are also possible. For example, a valid combination could be person.address('home').street[0].

2.2.1. Escape Characters

If your application uses System.Collection.Generic.IDictionary or XML to represent events, then event property names may themselves contain the dot ('.') character. The backslash ('\') character can be used to escape dot characters in property names, allowing a property name to contain dot characters.

For example, the EPL as shown below expects a property by name part1.part2 to exist on event type MyEvent:

select part1\.part2 from MyEvent

Sometimes your event properties may overlap with EPL language keywords. In this case you may use the backwards apostrophe ` character to escape the property name.

The next example assumes a Quote event that has a property by name order, while order is also a reserved keyword:

select `order` from Quote

2.3. Dynamic Event Properties

Dynamic (unchecked) properties are event properties that need not be known at statement compilation time. Such properties are resolved during runtime: they provide duck typing functionality.

The idea behind dynamic properties is that for a given underlying event representation we don't always know all properties in advance. An underlying event may have additional properties that are not known at statement compilation time, that we want to query on. The concept is especially useful for events that represent rich, object-oriented domain models.

The syntax of dynamic properties consists of the property name and a question mark. Indexed, mapped and nested properties can also be dynamic properties:

Table 2.3. Types of Event Properties

TypeSyntax
Dynamic Simple
name?
Dynamic Indexed
name[index]?
Dynamic Mapped
name('key')?
Dynamic Nested
name?.nestedPropertyName

Dynamic properties always return the System.Object type. Also, dynamic properties return a null value if the dynamic property does not exist on events processed at runtime.

As an example, consider an OrderEvent event that provides an "item" property. The "item" property is of type Object and holds a reference to an instance of either a Service or Product.

Assume that both Service and Product classes provide a property named "price". Via a dynamic property we can specify a query that obtains the price property from either object (Service or Product):

select item.price? from OrderEvent

As a second example, assume that the Service class contains a "serviceName" property that the Product class does not possess. The following query returns the value of the "serviceName" property for Service objects. It returns a null-value for Product objects that do not have the "serviceName" property:

select item.serviceName? from OrderEvent

Consider the case where OrderEvent has multiple implementation classes, some of which have a "timestamp" property. The next query returns the timestamp property of those implementations of the OrderEvent interface that feature the property:

select timestamp? from OrderEvent

The query as above returns a single column named "timestamp?" of type Object.

When dynamic properties are nested, then all properties under the dynamic property are also considered dynamic properties. In the below example the query asks for the "direction" property of the object returned by the "detail" dynamic property:

select detail?.direction from OrderEvent

Above is equivalent to:

select detail?.direction? from OrderEvent

The functions that are often useful in conjunction with dynamic properties are:

  • The cast function casts the value of a dynamic property (or the value of an expression) to a given type.

  • The exists function checks whether a dynamic property exists. It returns true if the event has a property of that name, or false if the property does not exist on that event.

  • The instanceof function checks whether the value of a dynamic property (or the value of an expression) is of any of the given types.

  • The typeof function returns the string type name of a dynamic property.

Dynamic event properties work with all event representations outlined next: Objects, Map-based and XML DOM-based events.

2.4. Fragment and Fragment Type

Sometimes an event can have properties that are itself events. Esper uses the term fragment and fragment type for such event pieces. The best example is a pattern that matches two or more events and the output event contains the matching events as fragments. In other words, output events can be a composite event that consists of further events, the fragments.

Fragments have the same metadata available as their enclosing composite events. The metadata for enclosing composite events contains information about which properties are fragments, or have a property value that can be represented as a fragment and therefore as an event itself.

Fragments and type metadata can allow your application to navigate composite events without the need for using the reflection API and reducing the coupling to the underlying event representation. The API is further described in Section 10.5, “Event and Event Type”.

2.5. Object Events

Object events are object instances that expose properties. While properties are the preferred method for exposing information through Esper, properties can also be exposed if an accessor-style or accessor-method is defined via configuration.

Esper supports event types that extend a superclass or implement one or more interfaces. Also, Esper event pattern and EPL statements can refer to interface classes and abstract classes.

Classes that represent events should be made immutable. As events are recordings of a state change or action that occurred in the past, the relevant event properties should not be changeable. However this is not a hard requirement and the Esper engine accepts events that are mutable as well.

The GetHashCode and Equals methods do not need to be implemented. The implementation of these methods by an event type does not affect the behavior of the engine in any way.

Please see Chapter 11, Configuration on options for naming event types represented by object event types. Types that do not follow standard property conventions, such as those that expose public fields, require additional configuration. Via configuration it is also possible to control case sensitivity in property name resolution. The relevant section in the chapter on configuration is Section 11.4.1.3, “Legacy Event Classes”.

2.5.1. Object Event Properties

As outlined earlier, the different property types are supported by the standard CLR properties specification, and some of which are uniquely supported by Esper:

  • Simple properties have a single value that may be retrieved. The underlying property type might be a CLR primitive (such as int, a simple object (such as a System.String), or a more complex object whose class is defined either by the CLR type structure, by the application, or by a type library included with the application.

  • Indexed - An indexed property stores an ordered collection of objects (all of the same type) that can be individually accessed by an integer-valued, non-negative index (or subscript).

  • Mapped - Esper considers any property that accepts a String-valued key a mapped property.

  • Nested - A nested property is a property that lives within another object which itself is a property of an event.

Assume there is an NewEmployeeEvent event type as shown below. The mapped and indexed properties in this example return objects but could also return CLR primitive types (such as Int32 or String). The Address object and Employee can themselves have properties that are nested within them, such as a street name in the Address object or a name of the employee in the Employee object.

public class NewEmployeeEvent {
	public String FirstName { get; }
	public Address GetAddress(String type);
	public Employee GetSubordinate(int index);
	public Employee[] AllSubordinates { get; }
}

Simple event properties require a read accessible type property returns the property value. This is standard convention for CLR-based applications and your specific language bindings will vary. In this example, the FirstName property returns the firstName event property of type String.

Indexed event properties require either one of the following getter-methods. A method that takes an integer-type key value and returns the property value, such as the GetSubordinate method, or a method that returns an array-type, or a class that implements IEnumerable. An example is the GetSubordinates getter method, which returns an array of Employee but could also return an Iterable. In an EPL or event pattern statement, indexed properties are accessed via the property[index] syntax.

Mapped event properties require a getter-method that takes a String-typed key value and returns the property value, such as the GetAddress method. In an EPL or event pattern statement, mapped properties are accessed via the property('key') syntax.

Nested event properties require a getter-method that returns the nesting object. The Address property and GetSubordinate method are mapped and indexed properties that return a nesting object. In an EPL or event pattern statement, nested properties are accessed via the property.NestedProperty syntax.

All event pattern and EPL statements allow the use of indexed, mapped and nested properties (or a combination of these) anywhere where one or more event property names are expected. The below example shows different combinations of indexed, mapped and nested properties in filters of event pattern expressions (each line is a separate EPL statement):

every NewEmployeeEvent(FirstName='myName')
every NewEmployeeEvent(Address('home').StreetName='Park Avenue')
every NewEmployeeEvent(Subordinate[0].Name='anotherName')
every NewEmployeeEvent(AllSubordinates[1].Name='thatName')
every NewEmployeeEvent(Subordinate[0].Address('home').StreetName='Water Street')

Similarly, the syntax can be used in EPL statements in all places where an event property name is expected, such as in select lists, where-clauses or join criteria.

select FirstName, Address('work'), Subordinate[0].Name, Subordinate[1].Name
from NewEmployeeEvent
where Address('work').StreetName = 'Park Ave'

2.5.2. Property Names

Property names are the result of a mixture of CLR standards and light method introspection of the type. Any properties that are defined via the CLR are automatically exposed. Additionally, any methods that are defined using the naming convention GetPropertyName is exposed as a property. In addition, Esper configuration provides a flag to turn off case-sensitive property names. A sample list of getter methods and property names is:

Table 2.4. Getter Methods and Property Names

MethodProperty NameExample
GetPrice()price
select Price from MyEvent
GetNAME()NAME
select NAME from MyEvent
GetItemDesc()itemDesc
select ItemDesc from MyEvent
GetQ()q
select Q from MyEvent
GetQN()QN
select QN from MyEvent
GetS()s
select S from MyEvent

It is important to note that Java and CLR property conventions are not compatible. In most cases, case conventions are opposite (e.g. Java uses camel case, whereas the CLR uses Pascal case).

2.5.3. Constants and Enumeration

Public static or const fields may also participate in expressions of all kinds, as this example shows:

select * from MyEvent where property=MyConstantClass.FIELD_VALUE

Event properties that are enumeration values can be compared by their enumeration value:

select * from MyEvent where enumProp=EnumClass.ENUM_VALUE_1

Alternatively, a static method may be employed on a class, such as the enumeration class 'EnumClass' as below:

select * from MyEvent where enumProp=EnumClass.valueOf('ENUM_VALUE_1')

If your application does not import, through configuration, the package that contains the enumeration type, then it must also specify the namespace of the type. Enumeration types that are inner classes must be qualified with + following conventions.

For example, the Color enumeration as an inner class to MyEvent in package org.myorg can be referenced as shown:

select * from MyEvent(enumProp=org.myorg.MyEvent+Color.GREEN).std:firstevent()

Instance methods may also be invoked on event instances by specifying a stream name, as shown below:

select myevent.computeSomething() as result from MyEvent as myevent

Chaining instance methods is supported as this example shows:

select myevent.GetComputerFor('books', 'movies').Calculate() as result 
from MyEvent as myevent

2.5.4. Parameterized Types

When your getter methods or accessor fields return a parameterized type, for example IEnumerable<MyEventData> for an indexed property or IDictionary<String, MyEventData> for a mapped property, then property expressions may refer to the properties available through the class that is the type parameter.

An example event that has properties that are parameterized types is:

public class NewEmployeeEvent {
  public String Name { get; }
  public IEnumerable<EducationHistory> Education { get; }
  public IDictionary<String, Address> Addresses { get; }
}

A sample of valid property expressions for this event is shown next:

select Name, Education, Education[0].Date, Addresses('home').Street
from NewEmployeeEvent

2.5.5. Known Limitations

Esper employs byte code generation for fast access to event properties. When byte code generation is unsuccessful, the engine logs a warning and uses CLR reflection to obtain property values instead.

A known limitation is that when an interface has an attribute of a particular type and the actual type returns a subclass of that attribute, the engine logs a warning and uses reflection for that property.

2.6. Map Events

2.6.1. Overview

Events can also be represented by objects that implement the System.Collection.IDictionary<String,Object> interface. Henceforth, we will refer to this type loosely as a DataMap. Event properties of DataMap events are the values in the map accessible through the indexer exposed by the IDictionary interface.

The Map event type is a comprehensive type system that can eliminate the need to use CLR types as event types, thereby making it easier to change types at runtime or generate type information from another source.

A given Map event type can have one or more supertypes that must also be Map event types. All properties available on any of the Map supertypes are available on the type itself. In addition, anywhere within EPL that an event type name of a Map supertype is used, any of its Map subtypes and their subtypes match that expression.

Your application can add properties to an existing Map event type during runtime using the configuration operation UpdateMapEventType. Properties may not be updated or deleted - properties can only be added, and nested properties can be added as well. The runtime configuration also allows removing Map event types and adding them back with new type information.

After your application configures a Map event type by providing a type name, the type name can be used when defining further Map event types by specifying the type name as a property type or an array property type.

One-to-Many relationships in Map event types are represented via arrays. A property in a Map event type may be an array of primitive, an array of object or an array of Map.

The engine can process DataMap events via the SendEvent(DataMap map, String eventTypeName) method on the EPRuntime interface. Entries in the Map represent event properties. Keys must be of type System.String for the engine to be able to look up event property names specified by pattern or EPL statements.

The engine does not validate Map event property names or values. Your application should ensure that objects passed in as event properties match the create schema property names and types, or the configured event type information when using runtime or static configuration.

2.6.2. Map Properties

Map event properties can be of any type. Map event properties that are application objects or that are of type DataMap (or arrays thereof) offer additional power:

  • Properties that are application objects can be queried via the nested, indexed, mapped and dynamic property syntax as outlined earlier.

  • Properties that are of type Map allow Maps to be nested arbitrarily deep and thus can be used to represent complex domain information. The nested, indexed, mapped and dynamic property syntax can be used to query Maps within Maps and arrays of Maps within Maps.

In order to use Map events, the event type name and property names and types must be made known to the engine via Configuration. Please see the examples in Section 11.4.2, “Events represented by System.Collection.Generic.IDictionary”.

The code snippet below creates and processes a Map event. It defines a CarLocationUpdateEvent event type first:

var mapEvent = new Dictionary<string,object>();
mapEvent["carId"] = carId;
mapEvent["direction"] = direction;
epRuntime.SendEvent(mapEvent, "CarLocUpdateEvent");

The CarLocUpdateEvent can now be used in a statement:

select carId from CarLocUpdateEvent.win:time(1 min) where direction = 1

The engine can also query objects as values in a Map event via the nested property syntax. Thus Map events can be used to aggregate multiple data structures into a single event and query the composite information in a convenient way. The example below demonstrates a Map event with a transaction and an account object.

var mapEvent = new Dictionary<string,object>();
mapEvent["txn"] = txn;
mapEvent["account"] = account;
epRuntime.SendEvent(mapEvent, "TxnEvent");

An example statement could look as follows.

select account.id, account.rate * txn.amount 
from TxnEvent.win:time(60 sec) 
group by account.id

2.6.3. Map Supertypes

Your Map event type may declare one or more supertypes when configuring the type at engine initialization time or at runtime through the administrative interface.

Supertypes of a Map event type must also be Map event types. All property names and types of a supertype are also available on a subtype and override such same-name properties of the subtype. In addition, anywhere within EPL that an event type name of a Map supertype is used, any of its Map subtypes also matches that expression (similar to the concept of interface).

This example assumes that the BaseUpdate event type has been declared and acts as a supertype to the AccountUpdate event type (both Map event types):

epService.EPAdministrator.GetConfiguration().
    AddEventType("AccountUpdate", accountUpdateDef, 
    new String[] {"BaseUpdate"});

Your application EPL statements may select BaseUpdate events and receive both BaseUpdate and AccountUpdate events, as well as any other subtypes of BaseUpdate and their subtypes.

// Receive BaseUpdate and any subtypes including subtypes of subtypes
select * from BaseUpdate

Your application Map event type may have multiple supertypes. The multiple inheritance hierarchy between Maps can be arbitrarily deep, however cyclic dependencies are not allowed. If using runtime configuration, supertypes must exist before a subtype to a supertype can be added.

See Section 11.4.2, “Events represented by System.Collection.Generic.IDictionary” for more information on configuring Map event types.

2.6.4. Advanced Map Property Types

2.6.4.1. Nested Properties

Strongly-typed nested Map-within-Map events can be used to build rich, type-safe event types on the fly. Use the AddEventType method on Configuration or ConfigurationOperations for initialization-time and runtime-time type definition.

Noteworthy points are:

  • Objects can appear as properties in Map-within-Map.

  • One may represent Map-within-Map and Map-Array within Map using the name of a previously registered Map event type.

  • There is no limit to the number of nesting levels.

  • Dynamic properties can be used to query Map-within-Map keys that may not be known in advance.

  • The engine returns a null value for properties for which the access path into the nested structure cannot be followed where map entries do not exist.

For demonstration, in this example our top-level event type is an AccountUpdate event, which has an UpdatedField structure as a property. Inside the UpdatedField structure the example defines various fields, as well as a property by name 'history' that holds a type UpdateHistory to represent the update history for the account. The code snippet to define the event type is thus:

IDictionary<String, Object> updatedFieldDef = new Dictionary<String, Object>();
updatedFieldDef["name"] = typeof(string);
updatedFieldDef["addressLine1"] = typeof(string);
updatedFieldDef["history"] = typeof(UpdateHistory);

IDictionary<String, Object> accountUpdateDef = new Dictionary<String, Object>();
accountUpdateDef["accountId"] = typeof(long);
accountUpdateDef["fields"] = updatedFieldDef;

epService.EPAdministrator.GetConfiguration().
    AddEventType("AccountUpdate", accountUpdateDef);

The next code snippet populates a sample event and sends the event into the engine:

IDictionary<String, Object> updatedField = new Dictionary<String, Object>();
updatedField["name"] = "Joe Doe";
updatedField["addressLine1"] = "40 Popular Street";
updatedField["history"] = new UpdateHistory();

IDictionary<String, Object> accountUpdate = new Dictionary<String, Object>();
accountUpdate["accountId"] = 10009901;
accountUpdate["fields"] = updatedField;

epService.EPRuntime.SendEvent(accountUpdate, "AccountUpdate");

Last, a sample query to interrogate AccountUpdate events is as follows:

select accountId, fields.name, fields.addressLine1, fields.history.lastUpdate
from AccountUpdate

Note that type information for nested maps is only available to the immediately selecting stream. For example, the second select-query does not work:

insert into MyStream select fields from NestedMapEvent
// this does not work ... instead select the individual fields in the insert-into statement
select fields.name from MyStream 

2.6.4.2. Map Event Type Properties

Your application may declare a Map event type for reuse within other Map event types or for one-to-many properties represented by an array of Maps.

This example declares a Map event type by name AmountCurrency with amount and currency properties:

IDictionary<String, Object> amountAndCurr = new Dictionary<String, Object>();
amountAndCurr["amount"] = typeof(double);
amountAndCurr["currency"] = typeof(string);

epService.EPAdministrator.GetConfiguration().
    AddEventType("AmountCurrency", amountAndCurr);

The AmountCurrency type is now available for use as a property type itself. Below code snippet declares OrderItem to hold an item number and AmountCurrency:

IDictionary<String, Object> orderItem = new Dictionary<String, Object>();
orderItem["itemNum"] = typeof(int);
orderItem["price"] = "AmountCurrency";    // The property type is the name itself

epService.EPAdministrator.GetConfiguration().
    AddEventType("OrderItem", orderItem);

2.6.4.3. One-to-Many Relationships

To model repeated properties within a Map, you may use arrays as properties in a Map. You may use an array of primitive types or an array of objects or an array of a previously declared Map event type.

When using a previously declared Map event type as an array property, the literal [] must be appended after the event type name.

This following example defines a Map event type by name Sale to hold array properties of the various types. It assumes a SalesPerson type exists and a Map event type by name OrderItem was declared:

Map<String, Object> sale = new HashMap<String, Object>();
sale["userids"] = typeof(int[]);
sale["salesPersons"] = typeof(SalesPerson[]);
sale["items"] = "OrderItem[]";	 // The property type is the name itself appended by []

epService.EPAdministrator.GetConfiguration().
    AddEventType("SaleEvent", sale);

The three properties that the above example declares are:

  • An integer array of user ids.

  • An array of SalesPerson objects.

  • An array of Maps for order items.

The next EPL statement is a sample query asking for property values held by arrays:

select userids[0], salesPersons[1].name, 
    items[1], items[1].price.amount from SaleEvent

2.7. System.Xml.XmlNode XML Events

Events can be represented as System.Xml.XmlNode instances and send into the engine via the SendEvent method on EPRuntime or via EventSender. Please note that configuration is required so the event type name and root element name is known. See Chapter 11, Configuration.

If a XML schema document (XSD file) can be made available as part of the configuration, then Esper can read the schema and appropriately present event type metadata and validate statements that use the event type and its properties. See Section 2.7.1, “Schema-Provided XML Events”.

When no XML schema document is provided, XML events can still be queried, however the return type and return values of property expressions are string-only and no event type metadata is available other then for explicitly configured properties. See Section 2.7.2, “No-Schema-Provided XML Events”.

In all cases Esper allows you to configure explicit XPath expressions as event properties. You can specify arbitrary XPath functions or expressions and provide a property name and type by which result values will be available for use in EPL statements. See Section 2.7.3, “Explicitly-Configured Properties”.

Nested, mapped and indexed event properties are also supported in expressions against System.Xml.XmlNode events. Thus XML trees can conveniently be interrogated via the property expression syntax.

Only one event type per root element name may be configured. The engine recognizes each event by its root element name or you may use EventSender to send events.

This section uses the following XML document as an example:

<?xml version="1.0" encoding="UTF-8"?>
<Sensor xmlns="SensorSchema">
  <ID>urn:epc:1:4.16.36</ID>
  <Observation Command="READ_PALLET_TAGS_ONLY">
    <ID>00000001</ID>
    <Tag>
      <ID>urn:epc:1:2.24.400</ID>
    </Tag>
    <Tag>
      <ID>urn:epc:1:2.24.401</ID>
    </Tag>
  </Observation>
</Sensor>

The schema for the example is:

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">

  <xs:element name="Sensor">
    <xs:complexType>
      <xs:sequence>
        <xs:element name="ID" type="xs:string"/>
        <xs:element ref="Observation" />
      </xs:sequence>
    </xs:complexType>
  </xs:element>

  <xs:element name="Observation">
    <xs:complexType>
      <xs:sequence>
        <xs:element name="ID" type="xs:string"/>
        <xs:element ref="Tag" maxOccurs="unbounded" />
      </xs:sequence>
      <xs:attribute name="Command" type="xs:string" use="required" />
    </xs:complexType>
  </xs:element>

  <xs:element name="Tag">
    <xs:complexType>
      <xs:sequence>
        <xs:element name="ID" type="xs:string"/>
      </xs:sequence>
    </xs:complexType>
  </xs:element>
</xs:schema>

2.7.1. Schema-Provided XML Events

If you have a XSD schema document available for your XML events, Esper can interrogate the schema. The benefits are:

  • New EPL statements that refer to event properties are validated against the types provided in the schema.

  • Event type metadata becomes available for retrieval as part of the EventType interface.

2.7.1.1. Getting Started

The engine reads a XSD schema file from an URL you provide. Make sure files imported by the XSD schema file can also be resolved.

The configuration accepts a schema URL. This is a sample code snippet to determine a schema URL from the resource manager:

URL schemaURL = ResourceManager.ResolveResourceURL("sensor.xsd");

Here is a sample use of the runtime configuration API, please see Chapter 11, Configuration for further examples.

epService = EPServiceProviderManager.GetDefaultProvider();
ConfigurationEventTypeXMLDOM sensorcfg = new ConfigurationEventTypeXMLDOM();
sensorcfg.RootElementName = "Sensor");
sensorcfg.SchemaResource = schemaURL.ToString();
epService.EPAdministrator.GetConfiguration()
    .AddEventType("SensorEvent", sensorcfg);

You must provide a root element name. This name is used to look up the event type for the SendEvent(System.Xml.XmlNode node) method. An EventSender is a useful alternative method for sending events if the type lookup based on the root or document element name is not desired.

After adding the event type, you may create statements and send events. Next is a sample statement:

select ID, Observation.Command, Observation.ID, 
  Observation.Tag[0].ID, Observation.Tag[1].ID
from SensorEvent

As you can see from the example above, property expressions can query property values held in the XML document's elements and attributes.

There are multiple ways to obtain a XML DOM document instance from a XML string. The next code snippet shows how to obtain a XML DOM System.Xml.XmlDocument instance:

XmlDocument doc = new XmlDocument();
doc.LoadXml(xml);

Send the System.Xml.XmlNode or Document object into the engine for processing:

epService.EPRuntime.SendEvent(doc);

2.7.1.2. Property Expressions and Namespaces

By default, property expressions such as Observation.Tag[0].ID are evaluated by a fast DOM-walker implementation provided by Esper. This DOM-walker implementation is not namespace-aware.

Should you require namespace-aware traversal of the DOM document, you must set the xpath-property-expr configuration option to true (default is false). This flag causes Esper to generate namespace-aware XPath expressions from each property expression instead of the DOM-walker, as described next. Setting the xpath-property-expr option to true requires that you also configure namespace prefixes as described below.

When matching up the property names with the XSD schema information, the engine determines whether the attribute or element provides values. The algorithm checks attribute names first followed by element names. It takes the first match to the specified property name.

2.7.1.3. Property Expression to XPath Rewrite

By setting the xpath-property-expr option the engine rewrites each property expression as an XPath expression, effectively handing the evaluation over to the underlying XPath implementation available from classpath. Most JVM have a built-in XPath implementation and there are also optimized, fast implementations such as Jaxen that can be used as well.

Set the xpath-property-expr option if you need namespace-aware document traversal, such as when your schema mixes several namespaces and element names are overlapping.

The below table samples several property expressions and the XPath expression generated for each, without namespace prefixes to keep the example simple:

Table 2.5. Property Expression to XPath Expression

Property ExpressionEquivalent XPath
Observeration.ID/Sensor/Observation/ID
Observeration.Command/Sensor/Observation/@Command
Observeration.Tag[0].ID/Sensor/Observation/Tag[position() = 1]/ID

For mapped properties that are specified via the syntax name('key'), the algorithm looks for an attribute by name id and generates a XPath expression as mapped[@id='key'].

Finally, here is an example that includes all different types of properties and their XPath expression equivalent in one property expression:

select nested.mapped('key').indexed[1].attribute from MyEvent

The equivalent XPath expression follows, this time including n0 as a sample namespace prefix:

/n0:rootelement/n0:nested/n0:mapped[@id='key']/n0:indexed[position() = 2]/@attribute

2.7.1.4. Array Properties

All elements that are unbound or have max occurs greater then 1 in the XSD schema are represented as indexed properties and require an index for resolution.

For example, the following is not a valid property expression in the sample Sensor document: Observeration.Tag.ID. As no index is provided for Tag, the property expression is not valid.

Repeated elements within a parent element in which the repeated element is a simple type also are represented as an array.

Consider the next XML document:

<item>
  <book sku="8800090">
    <author>Isaac Asimov</author>
    <author>Robert A Heinlein</author>
  </book>
</item>

Here, the result of the expression book.author is an array of type String and the result of book.author[0] is a String value.

2.7.1.5. Dynamic Properties

Dynamic properties are not validated against the XSD schema information and their result value is always System.Xml.XmlNode. You may use a user-defined function to process dynamic properties returning Node. As an alternative consider using an explicit property.

An example dynamic property is Origin?.ID which will look for an element by name Origin that contains an element or attribute node by name LocationCode:

select Origin?.LocationCode from SensorEvent

2.7.1.6. Transposing Properties

When providing a XSD document, the default configuration allows to transpose property values that are themselves complex elements, as defined in the XSD schema, into a new stream. This behavior can be controlled via the flag auto-fragment.

For example, consider the next query:

insert into ObservationStream
select ID, Observation from SensorEvent

The Observation as a property of the SensorEvent gets itself inserted into a new stream by name ObservationStream. The ObservationStream thus consists of a string-typed ID property and a complex-typed property named Observation, as described in the schema.

A further statement can use this stream to query:

select Observation.Command, Observation.Tag[0].ID from ObservationStream

Before continuing the discussion, here is an alternative syntax using the wildcard-select, that is also useful:

insert into TagListStream
select ID as sensorId, Observation.* from SensorEvent

The new TagListStream has a string-typed ID and Command property as well as an array of Tag properties that are complex types themselves as defined in the schema.

Next is a sample statement to query the new stream:

select sensorId, Command, Tag[0].ID from TagListStream

Please note the following limitations:

  • The XPath standard prescribes that XPath expressions against System.Xml.Node are evaluated against the owner document of the Node. Therefore XPath is not relative to the current node but absolute against each node's owner document. Since Esper does not create new document instances for transposed nodes, transposing properties is not possible when the xpath-property-expr flag is set.

  • Complex elements that have both simple element values and complex child elements are not transposed. This is to ensure their property value is not hidden. Use an explicit XPath expression to transpose such properties.

Esper automatically registers a new event type for transposed properties. It generates the type name of the new XML event type from the XML event type name and the property names used in the expression. The synposis is type_name.property_name[.property_name...]. The type name can be looked up, for example for use with EventSender or can be created in advance.

2.7.1.7. Event Sender

An IEventSender sends events into the engine for a given type, saving a type lookup based on element name.

This brief example sends an event via IEventSender:

EventSender sender = epRuntime.GetEventSender("SensorEvent");
sender.sendEvent(node);

The XML DOM event sender checks the root element name before processing the event. Use the event-sender-validates-root setting to disable validation. This forces the engine to process XML documents according to any predefined type without validation of the root element name.

2.7.2. No-Schema-Provided XML Events

Without a schema document a XML event may still be queried. However there are important differences in the metadata available without a schema document and therefore the property expression results. These differences are outlined below.

All property expressions against a XML type without schema are assumed valid. There is no validation of the property expression other then syntax validation. At runtime, property expressions return string-type values or null if the expression did not yield a matching element or attribute result.

When asked for property names or property metadata, a no-schema type returns empty array.

In all other aspects the type behaves the same as the schema-provided type described earlier.

2.7.3. Explicitly-Configured Properties

Regardless of whether or not you provide a XSD schema for the XML event type, you can always fall back to configuring explicit properties that are backed by XPath expressions.

For further documentation on XPath, please consult the XPath standard or other online material.

2.7.3.1. Simple Explicit Property

Shown below is an example configuration that adds an explicit property backed by a XPath expression and that defines namespace prefixes:

epService = EPServiceProviderManager.GetDefaultProvider();
ConfigurationEventTypeXMLDOM sensorcfg = new ConfigurationEventTypeXMLDOM();
sensorcfg.AddXPathProperty("countTags", "count(/ss:Sensor/ss:Observation/ss:Tag)", 
    XPathResultType.Number);
sensorcfg.AddNamespacePrefix("ss", "SensorSchema");
sensorcfg.RootElementName = "Sensor";
epService.EPAdministrator.GetConfiguration()
    .AddEventType("SensorEvent", sensorcfg);

The countTags property is now available for querying:

select countTags from SensorEvent

The XPath expression count(...) is a XPath built-in function that counts the number of nodes, for the example document the result is 2.

2.7.3.2. Explicit Property Casting and Parsing

Esper can parse or cast the result of your XPath expression to the desired type. Your property configuration provides the type to cast to, like this:

sensorcfg.AddXPathProperty("countTags", "count(/ss:Sensor/ss:Observation/ss:Tag)", 
    XPathResultType.Number, "int");

The type supplied to the property configuration must be one of the built-in types. Arrays of built-in type are also possible, requiring the XPathResultType.NodeSet type returned by your XPath expression, as follows:

sensorcfg.AddXPathProperty("idarray", "//ss:Tag/ss:ID", 
    XPathResultType.NodeSet, "String[]");

The XPath expression //ss:Tag/ss:ID returns all ID nodes under a Tag node, regardless of where in the node tree the element is located. For the example document the result is 2 array elements urn:epc:1:2.24.400 and urn:epc:1:2.24.40.

2.7.3.3. Node and Nodeset Explicit Property

An explicit property may return XPathResultType.Any or XPathResultType.NodeSet and can provide the event type name of a pre-configured event type for the property. The method name to add such properties is AddXPathPropertyFragment.

This code snippet adds two explicit properties and assigns an event type name for each property:

sensorcfg.AddXPathPropertyFragment("tagOne", "//ss:Tag[position() = 1]", 
    XPathResultType.Any, "TagEvent");
sensorcfg.AddXPathPropertyFragment("tagArray", "//ss:Tag", 
    XPathResultType.NodeSet, "TagEvent");

The configuration above references the TagEvent event type. This type must also be configured. Prefix the root element name with "//" to cause the lookup to search the nested schema elements for the definition of the type:

ConfigurationEventTypeXMLDOM tagcfg = new ConfigurationEventTypeXMLDOM();
tagcfg.RootElementName = "//Tag";
tagcfg.SchemaResource = schemaURL;
epAdministrator.GetConfiguration()
    .AddEventType("TagEvent", tagcfg);

The tagOne and tagArray properties are now ready for selection and transposing to further streams:

insert into TagOneStream select tagOne.* from SensorEvent

Select from the new stream:

select ID from TagOneStream

An example with indexed properties is shown next:

insert into TagArrayStream select tagArray as mytags from SensorEvent

Select from the new stream:

select mytags[0].ID from TagArrayStream

2.8. Additional Event Representations

Part of the extension and plug-in features of Esper is an event representation API. This set of classes allow an application to create new event types and event instances based on information available elsewhere, statically or dynamically at runtime when EPL statements are created. Please see Section 13.6, “Custom Event Representation” for details.

Creating a plug-in event representation can be useful when your application has existing types that carry event metadata and event property values and your application does not want to (or cannot) extract or transform such event metadata and event data into one of the built-in event representations (CLR objects, Map or XML DOM).

Further use of a plug-in event representation is to provide a faster or short-cut access path to event data. For example, access to event data stored in a XML format through the Streaming API for XML (StAX) is known to be very efficient. A plug-in event representation can also provide network lookup and dynamic resolution of event type and dynamic sourcing of event instances.

The chapter on Section 13.6, “Custom Event Representation” explains how to create your own custom event representation.

2.9. Updating, Merging and Versioning Events

To summarize, an event is an immutable record of a past occurrence of an action or state change, and event properties contain useful information about an event.

The length of time an event is of interest to the event processing engine (retention time) depends on your EPL statements, and especially the data window, pattern and output rate limiting clauses of your statements.

During the retention time of an event more information about the event may become available, such as additional properties or changes to existing properties. Esper provides three concepts for handling updates to events.

The first means to handle updating events is the update istream clause as further described in Section 4.22, “Updating an Insert Stream: the Update IStream Clause”. It is useful when you need to update events as they enter a stream, before events are evaluated by any particular consuming statement to that stream.

The second means to update events is the on-merge and on-update clauses, for use with named windows only, as further described in Section 4.17.9, “Triggered Upsert using the On-Merge Clause” and Section 4.17.7, “Updating Named Windows: the On Update clause”. On-merge is similar to the SQL merge clause and provides what is known as an "Upsert" operation: Update existing events or if no existing event(s) are found then insert a new event, all in one atomic operation provided by a single EPL statement. On-update can be used to update individual properties of events held in a named window.

The third means to handle updating events is the revision event types, for use with named windows only, as further described in Section 4.17.11, “Versioning and Revision Event Type Use with Named Windows”. With revision event types one can declare, via configuration only, multiple different event types and then have the engine present a merged event type that contains a superset of properties of all merged types, and have the engine merge events as they arrive without additional EPL statements.

Note that patterns do not reflect changes to past events. For the temporal nature of patterns, any changes to events that were observed in the past do not reflect upon current pattern state.

2.10. Coarse-Grained Events

Your application events may consist of fairly comprehensive, coarse-grained structures or documents. For example in business-to-business integration scenarios, XML documents or other event objects can be rich deeply-nested graphs of event properties.

To extract information from a coarse-grained event or to perform bulk operations on the rows of the property graph in an event, Esper provides a convenient syntax: When specifying a filter expression in a pattern or in a select clause, it may contain a contained-event selection syntax, as further described in Section 4.21, “Contained-Event Selection”.

2.11. Event Objects Populated by Insert Into

The insert into clause can populate plain-old object events and IDictionary events directly from the results of select clause expressions. Simply use the event type name as the stream name in the insert into clause as described in Section 4.10, “Merging Streams and Continuous Insertion: the Insert Into Clause”.

The column names specified in the select and insert into clause must match available writable properties in the event object to be populated (the target event type). The expression result types of any expressions in the select clause must also be compatible with the property types of the target event type.

Consider the following example statement:

insert into com.mycompany.NewEmployeeEvent 
select fname as firstName, lname as lastName from HRSystemEvent

The above example specifies the fully-qualified class name of NewEmployeeEvent. The engine instantianes NewEmployeeEvent for each result row and populates the firstName and lastName properties of each instance from the result of select clause expressions. The HRSystemEvent in the example is assumed to have lname and fname properties.

Note how the example uses the as-keyword to assign column names that match the property names of the NewEmployeeEvent target event. If the property names of the source and target events are the same, the as-keyword is not required.

The next example is an alternate form and specifies property names within the insert into clause instead. The example also assumes that NewEmployeeEvent has been defined or imported via configuration since it does not specify the event class package name:

insert into NewEmployeeEvent(firstName, lastName) 
select fname, lname from HRSystemEvent

Finally, this example populates HRSystemEvent events. The example populates the value of a type property where the event has the value 'NEW' and populates a new event object with the value 'HIRED', copying the fname and lname property values to the new event object:

insert into HRSystemEvent 
select fname, lname, 'HIRED' as type from HRSystemEvent(type='NEW')

The matching of the select or insert into-clause column names to target event type's property names is case-sensitive. It is allowed to only populate a subset of all available columns in the target event type. Wildcard (*) is also allowed and copies all fields of the events or multiple events in a join.

For CLR object events, your event class must provide writable properties. The event type should also provide a default constructor taking no parameters. If your event class does not have a default constructor, your application may configure a factory method via ConfigurationEventTypeLegacy.

The engine follows CLR standards in terms of widening, performing widening automatically in cases where widening type conversion is allowed without loss of precision, for both boxed and primitive types.

When inserting array-typed properties into a CLR or Map-type underlying event the event definition should declare the target property as an array.

Please note the following limitations:

  • Event types that utilize XML System.Xml.XmlNode underlying event objects cannot be target of an insert into clause.

Chapter 3. Processing Model

3.1. Introduction

The Esper processing model is continuous: Update listeners and/or subscribers to statements receive updated data as soon as the engine processes events for that statement, according to the statement's choice of event streams, views, filters and output rates.

As outlined in Chapter 10, API Reference the delegate for event handlers is com.espertech.esper.client.UpdateEventHandler. Implementations are invoked by the engine when results become available:

A second, strongly-typed and native, highly-performant method of result delivery is provided: A subscriber object is a direct binding of query results to an object. The object receives statement results via method invocation. The subscriber class need not implement an interface or extend a superclass. Please see Section 10.3.3, “Setting a Subscriber Object”.

The engine provides statement results to update listeners by placing results in com.espertech.esper.client.EventBean instances. A typical listener implementation queries the EventBean instances via getter methods to obtain the statement-generated results.

The Get method on the EventBean interface can be used to retrieve result columns by name. The property name supplied to the Get method can also be used to query nested, indexed or array properties of object graphs as discussed in more detail in Chapter 2, Event Representations and Section 10.5, “Event and Event Type”

The Underlying properties on the EventBean interface allows update listeners to obtain the underlying event object. For wildcard selects, the underlying event is the event object that was sent into the engine via the SendEvent method. For joins and select clauses with expressions, the underlying object implements System.Collections.Generic.IDictionary.

3.2. Insert Stream

In this section we look at the output of a very simple EPL statement. The statement selects an event stream without using a data window and without applying any filtering, as follows:

select * from Withdrawal

This statement selects all Withdrawal events. Every time the engine processes an event of type Withdrawal or any sub-type of Withdrawal, it invokes all update listeners, handing the new event to each of the statement's listeners.

The term insert stream denotes the new events arriving, and entering a data window or aggregation. The insert stream in this example is the stream of arriving Withdrawal events, and is posted to listeners as new events.

The diagram below shows a series of Withdrawal events 1 to 6 arriving over time. The number in parenthesis is the withdrawal amount, an event property that is used in the examples that discuss filtering.

Output example for a simple statement

Figure 3.1. Output example for a simple statement

The example statement above results in only new events and no old events posted by the engine to the statement's listeners.

3.3. Insert and Remove Stream

A length window instructs the engine to only keep the last N events for a stream. The next statement applies a length window onto the Withdrawal event stream. The statement serves to illustrate the concept of data window and events entering and leaving a data window:

select * from Withdrawal.win:length(5)

The size of this statement's length window is five events. The engine enters all arriving Withdrawal events into the length window. When the length window is full, the oldest Withdrawal event is pushed out the window. The engine indicates to listeners all events entering the window as new events, and all events leaving the window as old events.

While the term insert stream denotes new events arriving, the term remove stream denotes events leaving a data window, or changing aggregation values. In this example, the remove stream is the stream of Withdrawal events that leave the length window, and such events are posted to listeners as old events.

The next diagram illustrates how the length window contents change as events arrive and shows the events posted to an update listener.

Output example for a length window

Figure 3.2. Output example for a length window

As before, all arriving events are posted as new events to listeners. In addition, when event W1 leaves the length window on arrival of event W6, it is posted as an old event to listeners.

Similar to a length window, a time window also keeps the most recent events up to a given time period. A time window of 5 seconds, for example, keeps the last 5 seconds of events. As seconds pass, the time window actively pushes the oldest events out of the window resulting in one or more old events posted to update listeners.

Note: By default the engine only delivers the insert stream to listeners and observers. EPL supports optional istream, irstream and rstream keywords on select-clauses and on insert-into clauses to control which stream to deliver, see Section 4.3.7, “Selecting insert and remove stream events”. There is also a related, engine-wide configuration setting described in Section 11.4.14, “Engine Settings related to Stream Selection”.

3.4. Filters and Where-clauses

Filters to event streams allow filtering events out of a given stream before events enter a data window. The statement below shows a filter that selects Withdrawal events with an amount value of 200 or more.

select * from Withdrawal(amount>=200).win:length(5)

With the filter, any Withdrawal events that have an amount of less then 200 do not enter the length window and are therefore not passed to update listeners. Filters are discussed in more detail in Section 4.4.1, “Filter-based Event Streams” and Section 5.4, “Filter Expressions In Patterns”.

Output example for a statement with an event stream filter

Figure 3.3. Output example for a statement with an event stream filter

The where-clause and having-clause in statements eliminate potential result rows at a later stage in processing, after events have been processed into a statement's data window or other views.

The next statement applies a where-clause to Withdrawal events. Where-clauses are discussed in more detail in Section 4.5, “Specifying Search Conditions: the Where Clause”.

select * from Withdrawal.win:length(5) where amount >= 200

The where-clause applies to both new events and old events. As the diagram below shows, arriving events enter the window however only events that pass the where-clause are handed to update listeners. Also, as events leave the data window, only those events that pass the conditions in the where-clause are posted to listeners as old events.

Output example for a statement with where-clause

Figure 3.4. Output example for a statement with where-clause

The where-clause can contain complex conditions while event stream filters are more restrictive in the type of filters that can be specified. The next statement's where-clause applies the ceil function of the System.Math library type in the where clause. The insert-into clause makes the results of the first statement available to the second statement:

insert into WithdrawalFiltered select * from Withdrawal where Math.Ceiling(amount) >= 200
select * from WithdrawalFiltered

3.5. Time Windows

In this section we explain the output model of statements employing a time window view and a time batch view.

3.5.1. Time Window

A time window is a moving window extending to the specified time interval into the past based on the system time. Time windows enable us to limit the number of events considered by a query, as do length windows.

As a practical example, consider the need to determine all accounts where the average withdrawal amount per account for the last 4 seconds of withdrawals is greater then 1000. The statement to solve this problem is shown below.

select account, avg(amount) 
from Withdrawal.win:time(4 sec) 
group by account
having amount > 1000

The next diagram serves to illustrate the functioning of a time window. For the diagram, we assume a query that simply selects the event itself and does not group or filter events.

select * from Withdrawal.win:time(4 sec)

The diagram starts at a given time t and displays the contents of the time window at t + 4 and t + 5 seconds and so on.

Output example for a statement with a time window

Figure 3.5. Output example for a statement with a time window

The activity as illustrated by the diagram:

  1. At time t + 4 seconds an event W1 arrives and enters the time window. The engine reports the new event to update listeners.

  2. At time t + 5 seconds an event W2 arrives and enters the time window. The engine reports the new event to update listeners.

  3. At time t + 6.5 seconds an event W3 arrives and enters the time window. The engine reports the new event to update listeners.

  4. At time t + 8 seconds event W1 leaves the time window. The engine reports the event as an old event to update listeners.

3.5.2. Time Batch

The time batch view buffers events and releases them every specified time interval in one update. Time windows control the evaluation of events, as does the length batch window.

The next diagram serves to illustrate the functioning of a time batch view. For the diagram, we assume a simple query as below:

select * from Withdrawal.win:time_batch(4 sec)

The diagram starts at a given time t and displays the contents of the time window at t + 4 and t + 5 seconds and so on.

Output example for a statement with a time batch view

Figure 3.6. Output example for a statement with a time batch view

The activity as illustrated by the diagram:

  1. At time t + 1 seconds an event W1 arrives and enters the batch. No call to inform update listeners occurs.

  2. At time t + 3 seconds an event W2 arrives and enters the batch. No call to inform update listeners occurs.

  3. At time t + 4 seconds the engine processes the batched events and a starts a new batch. The engine reports events W1 and W2 to update listeners.

  4. At time t + 6.5 seconds an event W3 arrives and enters the batch. No call to inform update listeners occurs.

  5. At time t + 8 seconds the engine processes the batched events and a starts a new batch. The engine reports the event W3 as new data to update listeners. The engine reports the events W1 and W2 as old data (prior batch) to update listeners.

3.6. Batch Windows

The built-in data windows that act on batches of events are the win:time_batch and the win:length_batch views. The win:time_batch data window collects events arriving during a given time interval and posts collected events as a batch to listeners at the end of the time interval. The win:length_batch data window collects a given number of events and posts collected events as a batch to listeners when the given number of events has collected.

Related to batch data windows is output rate limiting. While batch data windows retain events the output clause offered by output rate limiting can control or stabilize the rate at which events are output, see Section 4.7, “Stabilizing and Controlling Output: the Output Clause”.

Let's look at how a time batch window may be used:

select account, amount from Withdrawal.win:time_batch(1 sec)

The above statement collects events arriving during a one-second interval, at the end of which the engine posts the collected events as new events (insert stream) to each listener. The engine posts the events collected during the prior batch as old events (remove stream). The engine starts posting events to listeners one second after it receives the first event and thereon.

For statements containing aggregation functions and/or a group by clause, the engine posts consolidated aggregation results for an event batch. For example, consider the following statement:

select sum(amount) as mysum from Withdrawal.win:time_batch(1 sec)

Note that output rate limiting also generates batches of events following the output model as discussed here.

3.7. Aggregation and Grouping

3.7.1. Insert and Remove Stream

Statements that aggregate events via aggregation functions also post remove stream events as aggregated values change.

Consider the following statement that alerts when 2 Withdrawal events have been received:

select count(*) as mycount from Withdrawal having count(*) = 2

When the engine encounters the second withdrawal event, the engine posts a new event to update listeners. The value of the "mycount" property on that new event is 2. Additionally, when the engine encounters the third Withdrawal event, it posts an old event to update listeners containing the prior value of the count, if specifing the rstream keyword in the select clause to select the remove stream. The value of the "mycount" property on that old event is also 2.

Note the statement above does not specify a data window and thereby counts all arriving events since statement start. The statement above retains no events and its memory allocation is only the aggregation state, i.e. a single long value to represent count(*).

The istream or rstream keyword can be used to eliminate either new events or old events posted to listeners. The next statement uses the istream keyword causing the engine to call the listener only once when the second Withdrawal event is received:

select istream count(*) as mycount from Withdrawal having count(*) = 2

3.7.2. Output for Aggregation and Group-By

Following SQL (Standard Query Language) standards for queries against relational databases, the presence or absence of aggregation functions and the presence or absence of the group by clause dictates the number of rows posted by the engine to listeners. The next sections outline the output model for batched events under aggregation and grouping. The examples also apply to data windows that don't batch events and post results continously as events arrive or leave data windows. The examples also apply to patterns providing events when a complete pattern matches.

In summary, as in SQL, if your query only selects aggregation values, the engine provides one row of aggregated values. It provides that row every time the aggregation is updated (insert stream), which is when events arrive or a batch of events gets processed, and when the events leave a data window or a new batch of events arrives. The remove stream then consists of prior aggregation values.

Also as in SQL, if your query selects non-aggregated values along with aggregation values in the select clause, the engine provides a row per event. The insert stream then consists of the aggregation values at the time the event arrives, while the remove stream is the aggregation value at the time the event leaves a data window, if any is defined in your query.

The documentation provides output examples for query types in Appendix A, Output Reference and Samples, and the next sections outlines each query type.

3.7.2.1. Un-aggregated and Un-grouped

An example statement for the un-aggregated and un-grouped case is as follows:

select * from Withdrawal.win:time_batch(1 sec)

At the end of a time interval, the engine posts to listeners one row for each event arriving during the time interval.

The appendix provides a complete example including input and output events over time at Section A.2, “Output for Un-aggregated and Un-grouped Queries”

3.7.2.2. Fully Aggregated and Un-grouped

If your statement only selects aggregation values and does not group, your statement may look as the example below:

select sum(amount) 
from Withdrawal.win:time_batch(1 sec)

At the end of a time interval, the engine posts to listeners a single row indicating the aggregation result. The aggregation result aggregates all events collected during the time interval.

The appendix provides a complete example including input and output events over time at Section A.3, “Output for Fully-aggregated and Un-grouped Queries”

3.7.2.3. Aggregated and Un-Grouped

If your statement selects non-aggregated properties and aggregation values, and does not group, your statement may be similar to this statement:

select account, sum(amount) 
from Withdrawal.win:time_batch(1 sec)

At the end of a time interval, the engine posts to listeners one row per event. The aggregation result aggregates all events collected during the time interval.

The appendix provides a complete example including input and output events over time at Section A.4, “Output for Aggregated and Un-grouped Queries”

3.7.2.4. Fully Aggregated and Grouped

If your statement selects aggregation values and all non-aggregated properties in the select clause are listed in the group by clause, then your statement may look similar to this example:

select account, sum(amount) 
from Withdrawal.win:time_batch(1 sec) 
group by account

At the end of a time interval, the engine posts to listeners one row per unique account number. The aggregation result aggregates per unique account.

The appendix provides a complete example including input and output events over time at Section A.5, “Output for Fully-aggregated and Grouped Queries”

3.7.2.5. Aggregated and Grouped

If your statement selects non-aggregated properties and aggregation values, and groups only some properties using the group by clause, your statement may look as below:

select account, accountName, sum(amount) 
from Withdrawal.win:time_batch(1 sec) 
group by account

At the end of a time interval, the engine posts to listeners one row per event. The aggregation result aggregates per unique account.

The appendix provides a complete example including input and output events over time at Section A.6, “Output for Aggregated and Grouped Queries”

3.8. Event Visibility and Current Time

An event sent by your application or generated by statements is visible to all other statements in the same engine instance. Similarly, current time (the time horizon) moves forward for all statements in the same engine instance. Please see the Chapter 10, API Reference chapter for how to send events and how time moves forward through system time or via simulated time, and the possible threading models.

Within an Esper engine instance you can additionally control event visibility and current time on a statement level, under the term isolated service as described in Section 10.9, “Service Isolation”.

An isolated service provides a dedicated execution environment for one or more statements. Events sent to an isolated service are visible only within that isolated service. In the isolated service you can move time forward at the pace and resolution desired without impacting other statements that reside in the engine runtime or other isolated services. You can move statements between the engine and an isolated service.

Chapter 4. EPL Reference: Clauses

4.1. EPL Introduction

The Event Processing Language (EPL) is a SQL-like language with SELECT, FROM, WHERE, GROUP BY, HAVING and ORDER BY clauses. Streams replace tables as the source of data with events replacing rows as the basic unit of data. Since events are composed of data, the SQL concepts of correlation through joins, filtering and aggregation through grouping can be effectively leveraged.

The INSERT INTO clause is recast as a means of forwarding events to other streams for further downstream processing. External data accessible through JDBC may be queried and joined with the stream data. Additional clauses such as the PATTERN and OUTPUT clauses are also available to provide the missing SQL language constructs specific to event processing.

The purpose of the UPDATE clause is to update event properties. Update takes place before an event applies to any selecting statements or pattern statements.

EPL statements are used to derive and aggregate information from one or more streams of events, and to join or merge event streams. This section outlines EPL syntax. It also outlines the built-in views, which are the building blocks for deriving and aggregating information from event streams.

EPL statements contain definitions of one or more views. Similar to tables in a SQL statement, views define the data available for querying and filtering. Some views represent windows over a stream of events. Other views derive statistics from event properties, group events or handle unique event property values. Views can be staggered onto each other to build a chain of views. The Esper engine makes sure that views are reused among EPL statements for efficiency.

The built-in set of views is:

  1. Data window views: win:length, win:length_batch, win:time, win:time_batch, win:time_length_batch, win:time_accum, win:ext_timed, ext:sort_window, ext:time_order, std:unique, std:groupwin, std:lastevent, std:firstevent, std:firstunique, win:firstlength, win:firsttime.

  2. Views that derive statistics: std:size, stat:uni, stat:linest, stat:correl, stat:weighted_avg.

EPL provides the concept of named window. Named windows are data windows that can be inserted-into and deleted-from by one or more statements, and that can queried by one or more statements. Named windows have a global character, being visible and shared across an engine instance beyond a single statement. Use the CREATE WINDOW clause to create named windows. Use the INSERT INTO clause to insert data into a named window, the ON DELETE clause to remove events from a named window, the ON UPDATE clause to update events held by a named window and the ON SELECT clause to perform a query triggered by a pattern or arriving event on a named window. Finally, the name of the named window can occur in a statement's FROM clause to query a named window or include the named window in a join or subquery.

EPL allows execution of on-demand (fire-and-forget, non-continuous, triggered by API) queries against named windows through the runtime API. The query engine automatically indexes named window data for fast access by ON SELECT/UPDATE/INSERT/DELETE without the need to create an index explicitly. For fast on-demand query execution via runtime API use the CREATE INDEX syntax to create an explicit index.

Use CREATE SCHEMA to declare an event type.

Variables can come in handy to parameterize statements and change parameters on-the-fly and in response to events. Variables can be used in an expression anywhere in a statement as well as in the output clause for dynamic control of output rates.

Esper can be extended by plugging-in custom developed views and aggregation functions.

4.2. EPL Syntax

EPL queries are created and stored in the engine, and publish results to listeners as events are received by the engine or timer events occur that match the criteria specified in the query. Events can also be obtained from running EPL queries via the GetSafeEnumerator and GetEnumerator methods that provide a pull-data API.

The select clause in an EPL query specifies the event properties or events to retrieve. The from clause in an EPL query specifies the event stream definitions and stream names to use. The where clause in an EPL query specifies search conditions that specify which event or event combination to search for. For example, the following statement returns the average price for IBM stock ticks in the last 30 seconds.

select avg(price) from StockTick.win:time(30 sec) where symbol='IBM'

EPL queries follow the below syntax. EPL queries can be simple queries or more complex queries. A simple select contains only a select clause and a single stream definition. Complex EPL queries can be build that feature a more elaborate select list utilizing expressions, may join multiple streams, may contain a where clause with search conditions and so on.

[insert into insert_into_def]
select select_list
from stream_def [as name] [, stream_def [as name]] [,...]
[where search_conditions]
[group by grouping_expression_list]
[having grouping_search_conditions]
[output output_specification]
[order by order_by_expression_list]
[limit num_rows]

4.2.1. Specifying Time Periods

Time-based windows as well as pattern observers and guards take a time period as a parameter. Time periods follow the syntax below.

time-period : [year-part] [month-part] [week-part] [day-part] [hour-part] 
      [minute-part] [seconds-part] [milliseconds-part]

year-part : (number|variable_name) ("years" | "year")
month-part : (number|variable_name) ("months" | "month")
week-part : (number|variable_name) ("weeks" | "week")
day-part : (number|variable_name) ("days" | "day")
hour-part : (number|variable_name) ("hours" | "hour")
minute-part : (number|variable_name) ("minutes" | "minute" | "min")
seconds-part : (number|variable_name) ("seconds" | "second" | "sec")
milliseconds-part : (number|variable_name) ("milliseconds" | "millisecond" | "msec")

Some examples of time periods are:

10 seconds
10 minutes 30 seconds
20 sec 100 msec
1 day 2 hours 20 minutes 15 seconds 110 milliseconds
0.5 minutes
1 year
1 year 1 month

Variable names and substitution parameters '?' for prepared statements are also allowed as part of a time period expression.

A unit in the month part is equivalent to 30 days.

4.2.2. Using Comments

Comments can appear anywhere in the EPL or pattern statement text where whitespace is allowed. Comments can be written in two ways: slash-slash (// ...) comments and slash-star (/* ... */) comments.

Slash-slash comments extend to the end of the line:

// This comment extends to the end of the line.
// Two forward slashes with no whitespace between them begin such comments.

select * from MyEvent  // this is a slash-slash comment

// All of this text together is a valid statement.

Slash-star comments can span multiple lines:

/* This comment is a "slash-star" comment that spans multiple lines.
 * It begins with the slash-star sequence with no space between the '/' and '*' characters.
 * By convention, subsequent lines can begin with a star and are aligned, but this is 
 * not required.
 */		
select * from MyEvent  /* this also works */

Comments styles can also be mixed:

select field1, // first comment
  /* second comment*/  field2
  from MyEvent

4.2.3. Reserved Keywords

Certain words such as select, delete or set are reserved and may not be used as identifiers. Please consult Appendix B, Reserved Keywords for the list of reserved keywords and permitted keywords.

Names of built-in functions and certain auxiliary keywords are permitted as event property names and in the rename syntax of the select clause. For example, count is acceptable.

Consider the example below, which assumes that 'last' is an event property of MyEvent:

// valid
select last, count(*) as count from MyEvent

This example shows an incorrect use of a reserved keyword:

// invalid
select insert from MyEvent

EPL offers an escape syntax for reserved keywords: Event properties as well as event or stream names may be escaped via the backwards apostrophe ` (ASCII 96) character.

The next example queries an event type by name Order (a reserved keyword) that provides a property by name insert (a reserved keyword):

// valid
select `insert` from `Order`

4.2.4. Escaping Strings

You may surround string values by either double-quotes (") or single-quotes ('). When your string constant in an EPL statement itself contains double quotes or single quotes, you must escape the quotes.

Double and single quotes may be escaped by the backslash (\) character or by unicode notation. Unicode 0027 is a single quote (') and 0022 is a double quote (").

The sample EPL below escapes the single quote in the string constant John's, and filters out order events where the name value matches:

select * from OrderEvent(name='John\'s')
// ...equivalent to...
select * from OrderEvent(name='John\u0027s')

The next EPL escapes the string constant Quote "Hello":

select * from OrderEvent(description like "Quote \"Hello\"")
// is equivalent to
select * from OrderEvent(description like "Quote \u0022Hello\u0022")

When building an escape string via the API, escape the backslash, as shown in below code snippet:

epService.getEPAdministrator().createEPL("select * from OrderEvent(name='John\\'s')");
// ... and for double quotes...
epService.getEPAdministrator().createEPL("select * from OrderEvent(
  description like \"Quote \\\"Hello\\\"\")");

4.2.5. Data Types

EPL honors all built-in primitive and boxed types.

EPL also follows CLR standards in terms of widening, performing widening automatically in cases where widening type conversion is allowed without loss of precision, for both boxed and primitive types:

  1. byte to short, int, long, float, double, decimal

  2. short to int, long, float, or double, decimal

  3. char to int, long, float, or double, decimal

  4. int to long, float, or double, decimal

  5. long to float or double, decimal

  6. float to double or decimal

  7. double to decimal

In cases where loss of precision is possible because of narrowing requirements, EPL compilation outputs a compilation error.

EPL supports casting via the cast function.

EPL returns double-type values for division regardless of operand type. EPL can also be configured to follow CLR rules for integer arithmetic instead as described in Section 11.4.18, “Engine Settings related to Expression Evaluation”.

Division by zero returns positive or negative infinity. Division by zero can be configured to return null instead.

4.2.5.1. Data Type of Constants

An EPL constant is a number or a character string that indicates a fixed value. Constants can be used as expressions in many EPL statements, including variable assignment and case-when statements. They can also be used as parameter values for many built-in objects and clauses. Constants are also called literals.

EPL supports the standard SQL constant notation as well as data type literals.

The following are types of EPL constants:

Table 4.1. Types of EPL constants

TypeDescriptionExamples
stringA single character to an unlimited number of characters. Valid delimiters are the single quote (') or double quote (").
select 'volume' as field1,
   "sleep" as field2, 
  "\u0041" as unicodeA
booleanA boolean value.
select true as field1, 
  false as field2
integerAn integer value (4 byte).
select 1 as field1, 
  -1 as field2, 
  1e2 as field3
longA long value (8 byte). Use the "L" or "l" (lowercase L) suffix.
select 1L as field1, 
  1l as field2
doubleA double-precision 64-bit IEEE 754 floating point.
select 1.67 as field1, 
  167e-2 as field2, 
  1.67d as field3
floatA single-precision 32-bit IEEE 754 floating point. Use the "f" suffix.
select 1.2f as field1, 
  1.2F as field2
byteA 8-bit signed two's complement integer.
select 0x10 as field1

EPL does not have a single-byte character data type for its literals. Single character literals are treated as string.

Internal byte representation and boundary values of constants follow the CLR standard.

4.2.6. Annotation

An annotation is an addition made to information in a statement. Esper provides certain built-in annotations for defining statement name, adding a statement description or for tagging statements such as for managing statements or directing statement output. Other then the built-in annotations, applications can provide their own annotation classes that the EPL compiler can populate.

An annotation is part of the statement text and precedes the EPL select or pattern statement. Annotations are therefore part of the EPL grammar. The syntax for annotations follows the host language (Java, .NET) annotation syntax:

@annotation_name [(annotation_parameters)]

An annotation consists of the annotation name and optional annotation parameters. The annotation_name is the simple class name or fully-qualified class name of the annotation class. The optional annotation_parameters are a list of key-value pairs following the syntax:

@annotation_name (attribute_name = attribute_value, [name=value, ...])

The attribute_name is an identifier that must match the attributes defined by the annotation class. An attribute_value is a constant of any of the primitive types or string, an array, an enumeration value or another (nested) annotation. Null values are not allowed as annotation attribute values. Enumeration values are supported in EPL statements and not support in statements created via the CreatePattern method.

Use the Attributes property of EPStatement to obtain annotations provided via statement text.

4.2.6.1. Application-Provided Annotations

Your application may provide its own annotation classes. The engine detects and populates annotation instances for application annotation classes.

To enable the engine to recognize application annotation classes, your annotation name must include the package name (i.e. be fully-qualified) or your engine configuration must import the annotation class or package via the configuration API.

For example, assume that your application defines an annotation in its application code as follows:

public @interface ProcessMonitor {
  String processName();
  boolean isLongRunning default false;
  int[] subProcessIds;
}

Shown next is an EPL statement text that utilizes the annotation class defined earlier:

@ProcessMonitor(processName='CreditApproval',
  isLongRunning=true, subProcessIds = {1, 2, 3} )
  
select count(*) from ProcessEvent(processId in (1, 2, 3).win:time(30)

Above example assumes the ProcessMonitor annotation class is imported via configuration XML or API. Here is an example API call to import annotations provided by a package com.mycompany.myannotations:

epService.getEPAdministrator().getConfiguration().addImport("com.mycompany.myannotations.*");

4.2.6.2. Built-In Annotations

The list of built-in EPL annotations is:

Table 4.2. Built-In EPL Annotations

NamePurpose and AttributesExample
Name

Provides a statement name. Attributes are:

value : Statement name.

@Name("MyStatementName")
Description

Provides a statement textual description. Attributes are:

value : Statement description.

@Description("A statement description 
is placed here.")
Tag

For tagging a statement with additional information. Attributes are:

name : Tag name.

value : Tag value.

@Tag(name="MyTagName", 
 value="MyTagValue")
Priority

Applicable when an event (or schedule) matches filter criteria for multiple statements: Defines the order of statement processing (requires an engine-level setting).

Attributes are:

value : priority value.

@Priority(10)
Drop

Applicable when an event (or schedule) matches filter criteria for multiple statements, drops the event after processing the statement (requires an engine-level setting).

No attributes.

@Drop
Hint

For providing one or more hints towards how the engine should execute a statement. Attributes are:

value : A comma-separated list of one or more keywords.

@Hint('ITERATE_ONLY')
Hook

Use this annotation to register one or more statement-specific hooks providing a hook type for each individual hook, such as for SQL parameter, column or row conversion.

Attributes are the hook type and the hook itself (usually a import or class name):

@Hook(type=HookType.SQLCOL, hook='MyDBTypeConvertor')

The following example statement text specifies some of the built-in annotations in combination:

@Name("RevenuePerCustomer")
@Description("Outputs revenue per customer considering all events encountered so far.")
@Tag(name="grouping", value="customer")

select customerId, sum(revenue) from CustomerRevenueEvent

4.2.6.3. @Name

Use the @Name EPL annotation to specify a statement name within the EPL statement itself, as an alternative to specifying the statement name via API.

If your application is also providing a statement name through the API, the statement name provided through the API overrides the annotation-provided statement name.

Example:

@Name("SecurityFilter1") select * from SecurityFilter(ip="127.0.0.1")

4.2.6.4. @Description

Use the @Description EPL annotation to add a statement textual description.

Example:

@Description('This statement filters localhost.') select * from SecurityFilter(ip="127.0.0.1")

4.2.6.5. @Tag

Use the @Tag EPL annotation to tag statements with name-value pairs, effectively adding a property to the statement. The attributes name and value are of type string.

Example:

@Tag(name='ip_sensitive', value='Y') 
@Tag(name='author', value='Jim')
select * from SecurityFilter(ip="127.0.0.1")

4.2.6.6. @Priority

This annotation only takes effect if the engine-level setting for prioritized execution is set via configuration, as described in Section 11.4.19, “Engine Settings related to Execution of Statements”.

Use the @Priority EPL annotation to tag statements with a priority value. The default priority value is zero (0) for all statements. When an event (or single timer execution) requires processing the event for multiple statements, processing begins with the highest priority statement and ends with the lowest-priority statement.

Example:

@Priority(10) select * from SecurityFilter(ip="127.0.0.1")

4.2.6.7. @Drop

This annotation only takes effect if the engine-level setting for prioritized execution is set via configuration, as described in Section 11.4.19, “Engine Settings related to Execution of Statements”.

Use the @Drop EPL annotation to tag statements that preempt all other same or lower-priority statements. When an event (or single timer execution) requires processing the event for multiple statements, processing begins with the highest priority statement and ends with the first statement marked with @Drop, which becomes the last statement to process that event.

Unless a different priority is specified, the statement with the @Drop EPL annotation executes at priority 1. Thereby @Drop alone is an effective means to remove events from a stream.

Example:

@Drop select * from SecurityFilter(ip="127.0.0.1")

4.2.6.8. @Hint

A hint can be used to provide tips for the engine to affect statement execution. Hints change performance or memory-use of a statement but generally do not change its output.

The string value of a Hint annotation contains a keyword or a comma-separated list of multiple keywords. Hint keywords are case-insensitive. A list of hints is available in Section 15.2.22, “Consider using Hints”.

Example:

@Hint('disable_reclaim_group') 
select ipaddress, count(*) from SecurityFilter.win:time(60 sec) group by ipaddress

4.2.6.9. @Hook

A hook is for attaching a callback to a statement.

The type value of a Hook annotation defines the type of hook and the hook value is an imported or fully-qualified class name providing the callback implementation.

4.3. Choosing Event Properties And Events: the Select Clause

The select clause is required in all EPL statements. The select clause can be used to select all properties via the wildcard *, or to specify a list of event properties and expressions. The select clause defines the event type (event property names and types) of the resulting events published by the statement, or pulled from the statement via the iterator methods.

The select clause also offers optional istream, irstream and rstream keywords to control whether input stream, remove stream or input and remove stream events are posted to UpdateEventHandler instances and observers to a statement. By default, the engine provides only the insert stream to listener and observers. See Section 11.4.14, “Engine Settings related to Stream Selection” on how to change the default.

The syntax for the select clause is summarized below.

select [istream | irstream | rstream] [distinct] * | expression_list ... 

The istream keyword is the default, and indicates that the engine only delivers insert stream events to listeners and observers. The irstream keyword indicates that the engine delivers both insert and remove stream. Finally, the rstream keyword tells the engine to deliver only the remove stream.

The distinct keyword outputs only unique rows depending on the column list you have specified after it. It must occur after the select and after the optional stream keywords, as described in more detail below.

4.3.1. Choosing all event properties: select *

The syntax for selecting all event properties in a stream is:

select * from stream_def

The following statement selects StockTick events for the last 30 seconds of IBM stock ticks.

select * from StockTick(symbol='IBM').win:time(30 sec)

You may well be asking: Why does the statement specify a time window here? First, the statement is meant to demonstrate the use of * wildcard. When the engine pushes statement results to your listener and as the statement does not select remove stream events via rstream keyword, the listener receives only new events and the time window could be left off. By adding the time window the pull API (iterator API or JDBC driver) returns the last 30 seconds of events.

The * wildcard and expressions can also be combined in a select clause. The combination selects all event properties and in addition the computed values as specified by any additional expressions that are part of the select clause. Here is an example that selects all properties of stock tick events plus a computed product of price and volume that the statement names 'pricevolume':

select *, price * volume as pricevolume from StockTick

When using wildcard (*), Esper does not actually copy your event properties out of your event or events. It simply wraps your native type in an EventBean interface. Your application has access to the underlying event object through the Underlying property and has access to the property values through the Get method.

In a join statement, using the select * syntax selects one event property per stream to hold the event for that stream. The property name is the stream name in the from clause.

4.3.2. Choosing specific event properties

To choose the particular event properties to return:

select event_property [, event_property] [, ...] from stream_def

The following statement simply selects the symbol and price properties of stock ticks, and the total volume for stock tick events in a 60-second time window.

select symbol, price, sum(volume) from StockTick(symbol='IBM').win:time(60 sec)

The following statement declares a further view onto the event stream of stock ticks: the univariate statistics view (stat:uni). The statement selects the properties that this view derives from the stream, for the last 100 events of IBM stock ticks in the length window.

select datapoints, total, average, variance, stddev, stddevpa
from StockTick(symbol='IBM').win:length(100).stat:uni(volume)

4.3.3. Expressions

The select clause can contain one or more expressions.

select expression [, expression] [, ...] from stream_def

The following statement selects the volume multiplied by price for a time batch of the last 30 seconds of stock tick events.

select volume * price from StockTick.win:time_batch(30 sec)

4.3.4. Renaming event properties

Event properties and expressions can be renamed using below syntax.

select [event_property | expression] as identifier [, ...]

The following statement selects volume multiplied by price and specifies the name volPrice for the resulting column.

select volume * price as volPrice from StockTick

Identifiers cannot contain the "." (dot) character, i.e. "vol.price" is not a valid identifier for the rename syntax.

4.3.5. Choosing event properties and events in a join

If your statement is joining multiple streams, your may specify property names that are unique among the joined streams, or use wildcard (*) as explained earlier.

In case the property name in your select or other clauses is not unique considering all joined streams, you will need to use the name of the stream as a prefix to the property.

This example is a join between the two streams StockTick and News, respectively named as 'tick' and 'news'. The example selects from the StockTick event the symbol value using the 'tick' stream name as a prefix:

select tick.symbol from StockTick.win:time(10) as tick, News.win:time(10) as news
where news.symbol = tick.symbol

Use the wildcard (*) selector in a join to generate a property for each stream, with the property value being the event itself. The output events of the statement below have two properties: the 'tick' property holds the StockTick event and the 'news' property holds the News event:

select * from StockTick.win:time(10) as tick, News.win:time(10) as news

The following syntax can also be used to specify what stream's properties to select:

select stream_name.* [as name] from ...

The selection of tick.* selects the StockTick stream events only:

select tick.* from StockTick.win:time(10) as tick, News.win:time(10) as news
where tick.symbol = news.symbol

The next example uses the as keyword to name each stream's joined events. This instructs the engine to create a property for each named event:

select tick.* as stocktick, news.* as news 
from StockTick.win:time(10) as tick, News.win:time(10) as news
where stock.symbol = news.symbol

The output events of the above example have two properties 'stocktick' and 'news' that are the StockTick and News events.

The stream name itself, as further described in Section 4.4.5, “Using the Stream Name”, may be used within expressions or alone.

This example passes events to a user-defined function named compute and also shows insert-into to populate an event stream of combined events:

insert into TickNewStream select tick, news, MyLib.compute(news, tick) as result
from StockTick.win:time(10) as tick, News.win:time(10) as news
where tick.symbol = news.symbol
// second statement that uses the TickNewStream stream
select tick.price, news.text, result from TickNewStream

In summary, the stream_name.* streamname wildcard syntax can be used to select a stream as the underlying event or as a property, but cannot appear within an expression. While the stream_name syntax (without wildcard) always selects a property (and not as an underlying event), and can occur anywhere within an expression.

4.3.6. Choosing event properties and events from a pattern

If your statement employs pattern expressions, then your pattern expression tags events with a tag name. Each tag name becomes available for use as a property in the select clause and all other clauses.

For example, here is a very simple pattern that matches on every StockTick event received within 30 seconds after start of the statement. The sample selects the symbol and price properties of the matching events:

select tick.symbol as symbol, tick.price as price
from pattern[every tick=StockTick where timer:within(10 sec)]

The use of the wildcard selector, as shown in the next statement, creates a property for each tagged event in the output. The next statement outputs events that hold a single 'tick' property whose value is the event itself:

select * from pattern[every tick=StockTick where timer:within(10 sec)]

You may also select the matching event itself using the tick.* syntax. The engine outputs the StockTick event itself to listeners:

select tick.* from pattern[every tick=StockTick where timer:within(10 sec)]

A tag name as specified in a pattern is a valid expression itself. This example uses the insert into clause to make available the events matched by a pattern to further statements:

// make a new stream of ticks and news available
insert into StockTickAndNews 
select tick, news from pattern [every tick=StockTick -> news=News(symbol=tick.symbol)]
// second statement to select from the stream of ticks and news
select tick.symbol, tick.price, news.text from StockTickAndNews

4.3.7. Selecting insert and remove stream events

The optional istream, irstream and rstream keywords in the select clause control the event streams posted to listeners and observers to a statement.

If neither keyword is specified, and in the default engine configuration, the engine posts only insert stream events via the NewEvents property of the UpdateEventArgs delivered to UpdateEventHandler instances listening to the statement. The engine does not post remove stream events, by default.

The insert stream consists of the events entering the respective window(s) or stream(s) or aggregations, while the remove stream consists of the events leaving the respective window(s) or the changed aggregation result. See Chapter 3, Processing Model for more information on insert and remove streams.

The engine posts remove stream events to the OldEvents property of the UpdateEventArgs delivered to UpdateEventHandler intsances only if the irstream keyword occurs in the select clause. This behavior can be changed via engine-wide configuration as described in Section 11.4.14, “Engine Settings related to Stream Selection”.

By specifying the istream keyword you can instruct the engine to only post insert stream events via the NewEvents property to the UpdateEventHandler. The engine will then not post any remove stream events, and the OldEvents property is always a null value.

By specifying the irstream keyword you can instruct the engine to post both insert stream and remove stream events.

By specifying the rstream keyword you can instruct the engine to only post remove stream events via the NewEvents property to the UpdateEventHandler. The engine will then not post any insert stream events, and the OldEvents property is also always a null value.

The following statement selects only the events that are leaving the 30 second time window.

select rstream * from StockTick.win:time(30 sec)

The istream and rstream keywords in the select clause are matched by same-name keywords available in the insert into clause. While the keywords in the select clause control the event stream posted to listeners to the statement, the same keywords in the insert into clause specify the event stream that the engine makes available to other statements.

4.3.8. Qualifying property names and stream names

Property or column names can optionally be qualified by a stream name and the provider URI. The syntax is:

[[provider_URI.]stream_name.]property_name

The provider_URI is the URI supplied to the EPServiceProviderManager class, or the string default for the default provider.

This example assumes the provider is the default provider:

select MyEvent.myProperty from MyEvent
// ... equivalent to ...
select default.MyEvent.myProperty from MyEvent

Stream names can also be qualified by the provider URI. The syntax is:

[provider_URI.]stream_name

The next example assumes a provider URI by name of Processor:

select Processor.MyEvent.myProperty from Processor.MyEvent

4.3.9. Select Distinct

The optional distinct keyword removes duplicate output events from output. The keyword must occur after the select keyword and after the optional irstream keyword.

The distinct keyword in your select instructs the engine to consolidate, at time of output, the output event(s) and remove output events with identical property values. Duplicate removal only takes place when two or more events are output together at any one time, therefore distinct is typically used with a batch data window, output rate limiting, on-demand queries, on-select or iterator pull API.

If two or more output event objects have same property values for all properties of the event, the distinct removes all but one duplicated event before outputting events to listeners. Indexed, nested and mapped properties are considered in the comparison, if present in the output event.

The next example outputs sensor ids of temperature sensor events, but only every 10 seconds and only unique sensor id values during the 10 seconds:

select distinct sensorId from TemperatureSensorEvent output every 10 seconds

Use distinct with wildcard (*) to remove duplicate output events considering all properties of an event.

This example statement outputs all distinct events either when 100 events arrive or when 10 seconds passed, whichever occurs first:

select distinct * from TemperatureSensorEvent.win:time_length_batch(10, 100)

When selecting nested, indexed, mapped or dynamic properties in a select clause with distinct, it is relevant to know that the comparison uses hash code and the Equals semantics.

4.4. Specifying Event Streams: the From Clause

The from clause is required in all EPL statements. It specifies one or more event streams or named windows. Each event stream or named window can optionally be given a name by means of the as keyword.

from stream_def [as name] [unidirectional] [retain-union | retain-intersection] 
    [, stream_def [as stream_name]] [, ...]

The event stream definition stream_def as shown in the syntax above can consists of either a filter-based event stream definition or a pattern-based event stream definition.

For joins and outer joins, specify two or more event streams. Joins between pattern-based and filter-based event streams are also supported. Joins and the unidirectional keyword are described in more detail in Section 4.11, “Joining Event Streams”.

Esper supports joins against relational databases for access to historical or reference data as explained in Section 4.15, “Accessing Relational Data via SQL”. Esper can also join results returned by an arbitrary method invocation, as discussed in Section 4.16, “Accessing Non-Relational Data via Method Invocation”.

The stream_name is an optional identifier assigned to the stream. The stream name can itself occur in any expression and provides access to the event itself from the named stream. Also, a stream name may be combined with a method name to invoke instance methods on events of that stream.

For all streams with the exception of historical sources your query may employ data window views as outlined below. The retain-intersection (the default) and retain-union keywords build a union or intersection of two or more data windows as described in Section 4.4.4, “Multiple Data Window Views”.

4.4.1. Filter-based Event Streams

The stream_def syntax for a filter-based event stream is as below:

event_stream_name [(filter_criteria)] [contained_selection] [.view_spec] [.view_spec] [...]

The event_stream_name is either the name of an event type or name of an event stream populated by an insert into statement or the name of a named window.

The filter_criteria is optional and consists of a list of expressions filtering the events of the event stream, within parenthesis after the event stream name.

The contained_selection is optional and is for use with coarse-grained events that have properties that are themselves one or more events, see Section 4.21, “Contained-Event Selection” for the synopsis and examples.

The view_spec are optional view specifications, which are combinable definitions for retaining events and for deriving information from events.

The following EPL statement shows event type, filter criteria and views combined in one statement. It selects all event properties for the last 100 events of IBM stock ticks for volume. In the example, the event type is the fully qualified type name org.esper.example.StockTick. The expression filters for events where the property symbol has a value of "IBM". The optional view specifications for deriving data from the StockTick events are a length window and a view for computing statistics on volume. The name for the event stream is "volumeStats".

select * from 
  org.esper.example.StockTick(symbol='IBM').win:length(100).stat:uni(volume) as volumeStats

Esper filters out events in an event stream as defined by filter criteria before it sends events to subsequent views. Thus, compared to search conditions in a where clause, filter criteria remove unneeded events early. In the above example, events with a symbol other then IBM do not enter the time window.

4.4.1.1. Specifying an Event Type

The simplest form of filter is a filter for events of a given type without any conditions on the event property values. This filter matches any event of that type regardless of the event's properties. The example below is such a filter.

select * from com.mypackage.myevents.RfidEvent

Instead of the fully-qualified type name any other event name can be mapped via Configuration to a type, making the resulting statement more readable:

select * from RfidEvent

Interfaces and superclasses are also supported as event types. In the below example IRfidReadable is an interface class.

select * from org.myorg.rfid.IRfidReadable

4.4.1.2. Specifying Filter Criteria

The filtering criteria to filter for events with certain event property values are placed within parenthesis after the event type name:

select * from RfidEvent(category="Perishable")

All expressions can be used in filters, including static methods that return a boolean value:

select * from com.mycompany.RfidEvent(MyRFIDLib.isInRange(x, y) or (x < 0 and y < 0))

Filter expressions can be separated via a single comma ','. The comma represents a logical AND between filter expressions:

select * from RfidEvent(zone=1, category=10)
...is equivalent to...
select * from RfidEvent(zone=1 and category=10)

The following operators are highly optimized through indexing and are the preferred means of filtering in high-volume event streams:

  • equals =

  • not equals !=

  • comparison operators < , > , >=, <=

  • ranges

    • use the between keyword for a closed range where both endpoints are included

    • use the in keyword and round () or square brackets [] to control how endpoints are included

    • for inverted ranges use the not keyword and the between or in keywords

  • list-of-values checks using the in keyword or the not in keywords followed by a comma-separated list of values

At compile time as well as at run time, the engine scans new filter expressions for sub-expressions that can be indexed. Indexing filter values to match event properties of incoming events enables the engine to match incoming events faster. The above list of operators represents the set of operators that the engine can best convert into indexes. The use of comma or logical and in filter expressions does not impact optimizations by the engine.

4.4.1.3. Filtering Ranges

Ranges come in the following 4 varieties. The use of round () or square [] bracket dictates whether an endpoint is included or excluded. The low point and the high-point of the range are separated by the colon : character.

  • Open ranges that contain neither endpoint (low:high)

  • Closed ranges that contain both endpoints [low:high]. The equivalent 'between' keyword also defines a closed range.

  • Half-open ranges that contain the low endpoint but not the high endpoint [low:high)

  • Half-closed ranges that contain the high endpoint but not the low endpoint (low:high]

The next statement shows a filter specifying a range for x and y values of RFID events. The range includes both endpoints therefore uses [] hard brackets.

mypackage.RfidEvent(x in [100:200], y in [0:100])

The between keyword is equivalent for closed ranges. The same filter using the between keyword is:

mypackage.RfidEvent(x between 100 and 200, y between 0 and 50)

The not keyword can be used to determine if a value falls outside a given range:

mypackage.RfidEvent(x not in [0:100])

The equivalent statement using the between keyword is:

mypackage.RfidEvent(x not between 0 and 100)

4.4.1.4. Filtering Sets of Values

The in keyword for filter criteria determines if a given value matches any value in a list of values.

In this example we are interested in RFID events where the category matches any of the given values:

mypackage.RfidEvent(category in ('Perishable', 'Container'))

By using the not in keywords we can filter events with a property value that does not match any of the values in a list of values:

mypackage.RfidEvent(category not in ('Household', 'Electrical'))

4.4.1.5. Filter Limitations

The following restrictions apply to filter criteria:

  • Range and comparison operators require the event property to be of a numeric type.

  • Aggregation functions are not allowed within filter expressions.

  • The prev previous event function and the prior prior event function cannot be used in filter expressions.

4.4.2. Pattern-based Event Streams

Event pattern expressions can also be used to specify one or more event streams in an EPL statement. For pattern-based event streams, the event stream definition stream_def consists of the keyword pattern and a pattern expression in brackets []. The syntax for an event stream definition using a pattern expression is below. As in filter-based event streams, an optional list of views that derive data from the stream can be supplied.

pattern [pattern_expression] [.view_spec] [.view_spec] [...]

The next statement specifies an event stream that consists of both stock tick events and trade events. The example tags stock tick events with the name "tick" and trade events with the name "trade".

select * from pattern [every tick=StockTickEvent or every trade=TradeEvent]

This statement generates an event every time the engine receives either one of the event types. The generated events resemble a map with "tick" and "trade" keys. For stock tick events, the "tick" key value is the underlying stock tick event, and the "trade" key value is a null value. For trade events, the "trade" key value is the underlying trade event, and the "tick" key value is a null value.

Lets further refine this statement adding a view the gives us the last 30 seconds of either stock tick or trade events. Lets also select prices and a price total.

select tick.price as tickPrice, trade.price as tradePrice, 
       sum(tick.price) + sum(trade.price) as total
  from pattern [every tick=StockTickEvent or every trade=TradeEvent].win:time(30 sec)

Note that in the statement above tickPrice and tradePrice can each be null values depending on the event processed. Therefore, an aggregation function such as sum(tick.price + trade.price)) would always return null values as either of the two price properties are always a null value for any event matching the pattern. Use the coalesce function to handle null values, for example: sum(coalesce(tick.price, 0) + coalesce(trade.price, 0)).

4.4.3. Specifying Views

Views are used to specify an expiry policy for events (data window views) and also to derive data. Views can be staggered onto each other. See the section Chapter 9, EPL Reference: Views on the views available that also outlines the different types of views: Data Window views and Derived-Value views.

Views can optionally take one or more parameters. These parameters are expressions themselves that may consist of any combination of variables, arithmetic, user-defined function or substitution parameters for prepared statements, for example.

The example statement below outputs a count per expressway for car location events (contains information about the location of a car on a highway) of the last 60 seconds:

select expressway, count(*) from CarLocEvent.win:time(60) 
group by expressway

The next example serves to show staggering of views. It uses the std:groupwin view to create a separate length window per car id:

select cardId, expressway, direction, segment, count(*) 
from CarLocEvent.std:groupwin(carId).win:length(4) 
group by carId, expressway, direction, segment

The first view std:groupwin(carId) groups car location events by car id. The second view win:length(4) keeps a length window of the 4 last events, with one separate length window for each car id. The example reports the number of events per car id and per expressway, direction and segment considering the last 4 events for each car id only.

Note that the group by syntax is generally preferable over std:groupwin for grouping information as it is SQL-compliant, easier to read and does not create a separate data window per group. The std:groupwin in above example creates a separate data window (length window in the example) per group, demonstrating staggering views.

When views are staggered onto each other as a chain of views, then the insert and remove stream received by each view is the insert and remove stream made available by the view (or stream) earlier in the chain.

The special keep-all view keeps all events: It does not provide a remove stream, i.e. events are not removed from the keep-all view unless by means of the on-delete syntax or by revision events.

4.4.4. Multiple Data Window Views

Data window views provide an expiry policy that indicates when to remove events from the data window, with the exception of the keep-all data window which has no expiry policy and the std:groupwin grouped-window view for allocating a new data window per group.

EPL allows the freedom to use multiple data window views onto a stream and thus combine expiry policies. Combining data windows into an intersection (the default) or a union can achieve a useful strategy for retaining events and expiring events that are no longer of interest. Named windows and the on-delete syntax provide an additional degree of freedom.

In order to combine two or more data window views there is no keyword required. The retain-intersection keyword is the default and the retain-union keyword may instead be provided for a stream.

The concept of union and intersection come from Set mathematics. In the language of Set mathematics, two sets A and B can be "added" together: The intersection of A and B is the set of all things which are members of both A and B, i.e. the members two sets have "in common". The union of A and B is the set of all things which are members of either A or B.

Use the retain-intersection (the default) keyword to retain an intersection of all events as defined by two or more data windows. All events removed from any of the intersected data windows are entered into the remove stream. This is the default behavior if neither retain keyword is specified.

Use the retain-union keyword to retain a union of all events as defined by two or more data windows. Only events removed from all data windows are entered into the remove stream.

As you can see, it is the remove stream that the combined multiple data windows provide which differs when retaining an intersection and retaining a union, the insert stream is the same to all data windows and their staggered views. Therefore, when combining batching data windows with further data windows, the insert stream still remains the insert stream of the set overall (not batched). Consider using output snapshot to obtain regular updates instead of combining batch and other data windows.

The next example statement totals the price of OrderEvent events in a union of the last 30 seconds and unique by product name:

select sum(price) from OrderEvent.win:time(30 sec).std:unique(productName) retain-union

In the above statement, all OrderEvent events that are either less then 30 seconds old or that are the last event for the product name are considered.

Here is an example statement totals the price of OrderEvent events in an intersection of the last 30 seconds and unique by product name:

select sum(price) from OrderEvent.win:time(30 sec).std:unique(productName) retain-intersection

In the above statement, only those OrderEvent events that are both less then 30 seconds old and are the last event for the product name are considered.

For advanced users and for backward compatibility, it is possible to configure Esper to allow multiple data window views without either of the retain keywords, as described in Section 11.4.11.2, “Configuring Multi-Expiry Policy Defaults”.

4.4.5. Using the Stream Name

Your from clause may assign a name to each stream. This assigned stream name can serve any of the following purposes.

First, the stream name can be used to disambiguate property names. The stream_name.property_name syntax uniquely identifies which property to select if property names overlap between streams. Here is an example:

select prod.productId, ord.productId from ProductEvent as prod, OrderEvent as ord

Second, the stream name can be used with a wildcard (*) character to select events in a join, or assign new names to the streams in a join:

// Select ProductEvent only
select prod.* from ProductEvent as prod, OrderEvent
// Assign column names 'product' and 'order' to each event
select prod.* as product, ord.* as order from ProductEvent as prod, OrderEvent as ord

Further, the stream name by itself can occur in any expression: The engine passes the event itself to that expression. For example, the engine passes the ProductEvent and the OrderEvent to the user-defined function 'checkOrder':

select prod.productId, MyFunc.checkOrder(prod, ord) 
from ProductEvent as prod, OrderEvent as ord

Last, you may invoke an instance method on each event of a stream, and pass parameters to the instance method as well. Instance method calls are allowed anywhere in an expression.

The next statement demonstrates this capability by invoking a method 'computeTotal' on OrderEvent events and a method 'getMultiplier' on ProductEvent events:

select ord.computeTotal(prod.getMultiplier()) from ProductEvent as prod, OrderEvent as ord

Instance methods may also be chained: Your EPL may invoke a method on the result returned by a method invocation.

Assume that your product event exposes a method getZone which returns a zone object. Assume that the Zone class declares a method checkZone. This example statement invokes a method chain:

select prod.getZone().checkZone("zone 1") from ProductEvent as prod

4.5. Specifying Search Conditions: the Where Clause

The where clause is an optional clause in EPL statements. Via the where clause event streams can be joined and events can be filtered.

Comparison operators =, < , > , >=, <=, !=, <>, is null, is not null and logical combinations via and and or are supported in the where clause. The where clause can also introduce join conditions as outlined in Section 4.11, “Joining Event Streams”. where clauses can also contain expressions. Some examples are listed below.

...where fraud.severity = 5 and amount > 500
...where (orderItem.orderId is null) or (orderItem.class != 10)		 
...where (orderItem.orderId = null) or (orderItem.class <> 10)		 
...where itemCount / packageCount > 10		 

4.6. Aggregates and grouping: the Group-by Clause and the Having Clause

4.6.1. Using aggregate functions

The aggregate functions are sum, avg, count, max, min, median, stddev, avedev. You can use aggregate functions to calculate and summarize data from event properties. For example, to find out the total price for all stock tick events in the last 30 seconds, type:

select sum(price) from StockTickEvent.win:time(30 sec)

Here is the syntax for aggregate functions:

aggregate_function( [all | distinct] expression)

You can apply aggregate functions to all events in an event stream window or other view, or to one or more groups of events. From each set of events to which an aggregate function is applied, Esper generates a single value.

Expression is usually an event property name. However it can also be a constant, function, or any combination of event property names, constants, and functions connected by arithmetic operators.

For example, to find out the average price for all stock tick events in the last 30 seconds if the price was doubled:

select avg(price * 2) from StockTickEvent.win:time(30 seconds)

You can use the optional keyword distinct with all aggregate functions to eliminate duplicate values before the aggregate function is applied. The optional keyword all which performs the operation on all events is the default.

You can use aggregation functions in a select clause and in a having clause. You cannot use aggregate functions in a where clause, but you can use the where clause to restrict the events to which the aggregate is applied. The next query computes the average and sum of the price of stock tick events for the symbol IBM only, for the last 10 stock tick events regardless of their symbol.

select 'IBM stats' as title, avg(price) as avgPrice, sum(price) as sumPrice
from StockTickEvent.win:length(10)
where symbol='IBM'

In the above example the length window of 10 elements is not affected by the where clause, i.e. all events enter and leave the length window regardless of their symbol. If we only care about the last 10 IBM events, we need to add filter criteria as below.

select 'IBM stats' as title, avg(price) as avgPrice, sum(price) as sumPrice
from StockTickEvent(symbol='IBM').win:length(10)
where symbol='IBM'

You can use aggregate functions with any type of event property or expression, with the following exceptions:

  1. You can use sum, avg, median, stddev, avedev with numeric event properties only

Esper ignores any null values returned by the event property or expression on which the aggregate function is operating, except for the count(*) function, which counts null values as well. All aggregate functions return null if the data set contains no events, or if all events in the data set contain only null values for the aggregated expression.

4.6.2. Organizing statement results into groups: the Group-by clause

The group by clause is optional in all EPL statements. The group by clause divides the output of an EPL statement into groups. You can group by one or more event property names, or by the result of computed expressions. When used with aggregate functions, group by retrieves the calculations in each subgroup. You can use group by without aggregate functions, but generally that can produce confusing results.

For example, the below statement returns the total price per symbol for all stock tick events in the last 30 seconds:

select symbol, sum(price) from StockTickEvent.win:time(30 sec) group by symbol

The syntax of the group by clause is:

group by aggregate_free_expression [, aggregate_free_expression] [, ...]

Esper places the following restrictions on expressions in the group by clause:

  1. Expressions in the group by cannot contain aggregate functions

  2. Event properties that are used within aggregate functions in the select clause cannot also be used in a group by expression

  3. When grouping an unbound stream, i.e. no data window is specified onto the stream providing groups, or when using output rate limiting with the ALL keyword, you should ensure your group-by expression does not return an unlimited number of values. If, for example, your group-by expression is a fine-grained timestamp, group state that accumulates for an unlimited number of groups potentially reduces available memory significantly. Use a @Hint as described below to instruct the engine when to discard group state.

You can list more then one expression in the group by clause to nest groups. Once the sets are established with group by the aggregation functions are applied. This statement posts the median volume for all stock tick events in the last 30 seconds per symbol and tick data feed. Esper posts one event for each group to statement listeners:

select symbol, tickDataFeed, median(volume) 
from StockTickEvent.win:time(30 sec) 
group by symbol, tickDataFeed

In the statement above the event properties in the select list (symbol, tickDataFeed) are also listed in the group by clause. The statement thus follows the SQL standard which prescribes that non-aggregated event properties in the select list must match the group by columns.

Esper also supports statements in which one or more event properties in the select list are not listed in the group by clause. The statement below demonstrates this case. It calculates the standard deviation for the last 30 seconds of stock ticks aggregating by symbol and posting for each event the symbol, tickDataFeed and the standard deviation on price.

select symbol, tickDataFeed, stddev(price) from StockTickEvent.win:time(30 sec) group by symbol

The above example still aggregates the price event property based on the symbol, but produces one event per incoming event, not one event per group.

Additionally, Esper supports statements in which one or more event properties in the group by clause are not listed in the select list. This is an example that calculates the mean deviation per symbol and tickDataFeed and posts one event per group with symbol and mean deviation of price in the generated events. Since tickDataFeed is not in the posted results, this can potentially be confusing.

select symbol, avedev(price) 
from StockTickEvent.win:time(30 sec) 
group by symbol, tickDataFeed

Expressions are also allowed in the group by list:

select symbol * price, count(*) from StockTickEvent.win:time(30 sec) group by symbol * price

If the group by expression resulted in a null value, the null value becomes its own group. All null values are aggregated into the same group. If you are using the count(expression) aggregate function which does not count null values, the count returns zero if only null values are encountered.

You can use a where clause in a statement with group by. Events that do not satisfy the conditions in the where clause are eliminated before any grouping is done. For example, the statement below posts the number of stock ticks in the last 30 seconds with a volume larger then 100, posting one event per group (symbol).

select symbol, count(*) from StockTickEvent.win:time(30 sec) where volume > 100 group by symbol
4.6.2.1. Hints Pertaining to Group-By

The Esper engine reclaims aggregation state agressively when it determines that a group has no data points, based on the data in the data windows. When your application data creates a large number of groups with a small or zero number of data points then performance may suffer as state is reclaimed and created anew. Esper provides the @Hint('disable_reclaim_group') hint that you can specify as part of an EPL statement text to avoid group reclaim.

When aggregating values over an unbound stream (i.e. no data window is specified onto the stream) and when your group-by expression returns an unlimited number of values, for example when a timestamp expression is used, then please note the next hint.

A sample statement that aggregates stock tick events by timestamp, assuming the event type offers a property by name timestamp that, reflects time in high resolution, for example arrival or system time:

// Note the below statement could lead to an out-of-memory problem:
select symbol, sum(price) from StockTickEvent group by timestamp

As the engine has no means of detecting when aggregation state (sums per symbol) can be discarded, you may use the following hints to control aggregation state lifetime.

The @Hint("reclaim_group_aged=age_in_seconds") hint instructs the engine to discard aggregation state that has not been updated for age_in_seconds seconds.

The optional @Hint("reclaim_group_freq=sweep_frequency_in_seconds") can be used in addition to control the frequency at which the engine sweeps aggregation state to determine aggregation state age and remove state that is older then age_in_seconds seconds. If the hint is not specified, the frequency defaults to the same value as age_in_seconds.

The updated sample statement with both hints:

// Instruct engine to remove state older then 10 seconds and sweep every 5 seconds
@Hint('reclaim_group_aged=10,reclaim_group_freq=5')
select symbol, sum(price) from StockTickEvent group by timestamp

Variables may also be used to provide values for age_in_seconds and sweep_frequency_in_seconds.

This example statement uses a variable named varAge to control how long aggregation state remains in memory, and the engine defaults the sweep frequency to the same value as the variable provides:

@Hint('reclaim_group_aged=varAge')
select symbol, sum(price) from StockTickEvent group by timestamp

4.6.3. Selecting groups of events: the Having clause

Use the having clause to pass or reject events defined by the group-by clause. The having clause sets conditions for the group by clause in the same way where sets conditions for the select clause, except where cannot include aggregate functions, while having often does.

This statement is an example of a having clause with an aggregate function. It posts the total price per symbol for the last 30 seconds of stock tick events for only those symbols in which the total price exceeds 1000. The having clause eliminates all symbols where the total price is equal or less then 1000.

select symbol, sum(price) 
from StockTickEvent.win:time(30 sec) 
group by symbol 
having sum(price) > 1000

To include more then one condition in the having clause combine the conditions with and, or or not. This is shown in the statement below which selects only groups with a total price greater then 1000 and an average volume less then 500.

select symbol, sum(price), avg(volume)
from StockTickEvent.win:time(30 sec) 
group by symbol 
having sum(price) > 1000 and avg(volume) < 500

A statement with the having clause should also have a group by clause. If you omit group-by, all the events not excluded by the where clause return as a single group. In that case having acts like a where except that having can have aggregate functions.

The having clause can also be used without group by clause as the below example shows. The example below posts events where the price is less then the current running average price of all stock tick events in the last 30 seconds.

select symbol, price, avg(price) 
from StockTickEvent.win:time(30 sec) 
having price < avg(price)

4.6.4. How the stream filter, Where, Group By and Having clauses interact

When you include filters, the where condition, the group by clause and the having condition in an EPL statement the sequence in which each clause affects events determines the final result:

  1. The event stream's filter condition, if present, dictates which events enter a window (if one is used). The filter discards any events not meeting filter criteria.

  2. The where clause excludes events that do not meet its search condition.

  3. Aggregate functions in the select list calculate summary values for each group.

  4. The having clause excludes events from the final results that do not meet its search condition.

The following query illustrates the use of filter, where, group by and having clauses in one statement with a select clause containing an aggregate function.

select tickDataFeed, stddev(price)
from StockTickEvent(symbol='IBM').win:length(10) 
where volume > 1000
group by tickDataFeed 
having stddev(price) > 0.8

Esper filters events using the filter criteria for the event stream StockTickEvent. In the example above only events with symbol IBM enter the length window over the last 10 events, all other events are simply discarded. The where clause removes any events posted by the length window (events entering the window and event leaving the window) that do not match the condition of volume greater then 1000. Remaining events are applied to the stddev standard deviation aggregate function for each tick data feed as specified in the group by clause. Each tickDataFeed value generates one event. Esper applies the having clause and only lets events pass for tickDataFeed groups with a standard deviation of price greater then 0.8.

4.6.5. Comparing the Group By clause and the std:groupwin view

The group by clause as well as the built-in std:groupwin view are similar in their ability to group events. This section explains the key differences in their behavior and use.

The group by clause works together with aggregation functions in your statement to produce an aggregation result per group. In greater detail, this means that when a new event arrives, the engine applies the expressions in the group by clause to determine a grouping key. If the engine has not encountered that grouping key before (a new group), the engine creates a set of new aggregation results for that grouping key and performs the aggregation changing that new set of aggregation results. If the grouping key points to an existing set of prior aggregation results (an existing group), the engine performs the aggregation changing the prior set of aggregation results for that group.

The std:groupwin view is a built-in view that also groups events. The view is described in greater detail in Section 9.2.2, “Grouped Data Window (std:groupwin)”. Its primary use is to create a separate data window per group, or more generally to create separate instances of all its sub-views for each grouping key encountered.

The next example shows two queries that produce equivalent results. The query using the group by clause is generally preferable as is easier to read. The second form introduces the stat:uni view which computes univariate statistics for a given property:

select symbol, avg(price) from StockTickEvent group by symbol
// ... is equivalent to ...
select symbol, average from StockTickEvent.std:groupwin(symbol).stat:uni(price)

The next example shows two queries that are NOT equivalent as the length window is ungrouped in the first query, and grouped in the second query:

select symbol, sum(price) from StockTickEvent.win:length(10) group by symbol
// ... NOT equivalent to ...
select symbol, sum(price) from StockTickEvent.std:groupwin(symbol).win:length(10)

The key difference between the two statements is that in the first statement the length window is ungrouped and applies to all events regardless of group. While in the second query each group gets its own instance of a length window. For example, in the second query events arriving for symbol "ABC" get a length window of 10 events, and events arriving for symbol "DEF" get their own length window of 10 events.

4.7. Stabilizing and Controlling Output: the Output Clause

4.7.1. Output Clause Options

The output clause is optional in Esper and is used to control or stabilize the rate at which events are output and to suppress output events. The EPL language provides for several different ways to control output rate.

Here is the syntax for the output clause that specifies a rate in time interval or number of events:

output [after suppression_def] 
  [[all | first | last | snapshot] every output_rate [seconds | events]]

An alternate syntax specifies the time period between output as outlined in Section 4.2.1, “Specifying Time Periods” :

output [after suppression_def] 
  [[all | first | last | snapshot] every time_period]

A crontab-like schedule can also be specified. The schedule parameters follow the pattern observer parameters and are further described in Section 5.6.2.2, “timer:at” :

output [after suppression_def] 
  [[all | first | last | snapshot] at 
   (minutes, hours, days of month, months, days of week [, seconds])]

Last, output can be controlled by an expression that may contain variables, user-defined functions and information about the number of collected events. Output that is controlled by an expression is discussed in detail below.

The after keyword and suppression_def can appear alone or together with further output conditions and suppresses output events.

For example, the following statement outputs, every 60 seconds, the total price for all orders in the 30-minute time window:

select sum(price) from OrderEvent.win:time(30 min) output snapshot every 60 seconds

The all keyword is the default and specifies that all events in a batch should be output, each incoming row in the batch producing an output row. Note that for statements that group via the group by clause, the all keyword provides special behavior as below.

The first keyword specifies that only the first event in an output batch is to be output. Using the first keyword instructs the engine to output the first matching event as soon as it arrives, and then ignores matching events for the time interval or number of events specified. After the time interval elapsed, or the number of matching events has been reached, the next first matching event is output again and the following interval the engine again ignores matching events. For statements that group via the group by clause, the first keywords provides special behavior as below.

The last keyword specifies to only output the last event at the end of the given time interval or after the given number of matching events have been accumulated. Again, for statements that group via the group by clause the last keyword provides special behavior as below.

The snapshot keyword indicates that the engine output current computation results considering all events as per views specified and/or current aggregation results. While the other keywords control how a batch of events between output intervals is being considered, the snapshot keyword outputs all current state of a statement independent of the last batch. Its output is equivalent to the GetEnumerator method provided by a statement. The snapshot keyword requires a data window declaration if not used within a join and outputs only the last event if used without a data window.

The output_rate is the frequency at which the engine outputs events. It can be specified in terms of time or number of events. The value can be a number to denote a fixed output rate, or the name of a variable whose value is the output rate. By means of a variable the output rate can be controlled externally and changed dynamically at runtime.

Please consult the Appendix A, Output Reference and Samples for detailed information on insert and remove stream output for the various output clause keywords.

The time interval can also be specified in terms of minutes; the following statement is identical to the first one.

select * from StockTickEvent.win:length(5) output every 1.5 minutes

A second way that output can be stabilized is by batching events until a certain number of events have been collected. The next statement only outputs when either 5 (or more) new or 5 (or more) old events have been batched.

select * from StockTickEvent.win:time(30 sec) output every 5 events

Additionally, event output can be further modified by the optional last keyword, which causes output of only the last event to arrive into an output batch.

select * from StockTickEvent.win:time(30 sec) output last every 5 events

Using the first keyword you can be notified at the start of the interval. The allows to watch for situations such as a rate falling below a threshold and only be informed every now and again after the specified output interval, but be informed the moment it first happens.

select * from TickRate.win:time(30 seconds) where rate<100 output first every 60 seconds

A sample statement using the Unix "crontab"-command schedule is shown next. See Section 5.6.2.2, “timer:at” for details on schedule syntax. Here, output occurs every 15 minutes from 8am to 5:45pm (hours 8 to 17 at 0, 15, 30 and 45 minutes past the hour):

select symbol, sum(price) from StockTickEvent group by symbol output at (*/15, 8:17, *, *, *)

4.7.1.1. Controlling Output Using an Expression

Output can also be controlled by an expression that may check variable values, use user-defined functions and query built-in properties that provide additional information. The synopsis is as follows:

output [after suppression_def] 
  [[all | first | last | snapshot] when trigger_expression 
    [then set variable_name = assign_expression [, variable_name = assign_expression [,...]]]

The when keyword must be followed by a trigger expression returning a boolean value of true or false, indicating whether to output. Use the optional then keyword to change variable values after the trigger expression evaluates to true. An assignment expression assigns a new value to variable(s).

Lets consider an example. The next statement assumes that your application has defined a variable by name OutputTriggerVar of boolean type. The statement outputs rows only when the OutputTriggerVar variable has a boolean value of true:

select sum(price) from StockTickEvent output when OutputTriggerVar = true

The engine evaluates the trigger expression when streams and data views post one or more insert or remove stream events after considering the where clause, if present. It also evaluates the trigger expression when any of the variables used in the trigger expression, if any, changes value. Thus output occurs as follows:

  1. When there are insert or remove stream events and the when trigger expression evaluates to true, the engine outputs the resulting rows.

  2. When any of the variables in the when trigger expression changes value, the engine evaluates the expression and outputs results. Result output occurs within the minimum time interval of timer resolution (100 milliseconds).

By adding a then part to the EPL, we can reset any variables after the trigger expression evaluated to true:

select sum(price) from StockTickEvent 
  output when OutputTriggerVar = true  
  then set OutputTriggerVar = false

Expressions in the when and then may, for example, use variables, user defined functions or any of the built-in named properties that are described in the below list.

The following built-in properties are available for use:

Table 4.3. Built-In Properties for Use with Output When

Built-In Property NameDescription
last_output_timestampTimestamp when the last output occurred for the statement; Initially set to time of statement creation
count_insertNumber of insert stream events
count_removeNumber of remove stream events

The values provided by count_insert and count_remove are non-continues: The number returned for these properties may 'jump' up rather then count up by 1. The counts reset to zero upon output.

The following restrictions apply to expressions used in the output rate clause:

  • Event property names cannot be used in the output clause.

  • Aggregation functions cannot be used in the output clause.

  • The prev previous event function and the prior prior event function cannot be used in the output clause.

4.7.1.2. Suppressing Output With After

The after keyword and its time period or number of events parameters is optional and can occur after the output keyword, either alone or with output conditions as listed above.

The synopsis of after is as follows:

output after time_period | number events [...]

When using after either alone or together with further output conditions, the engine discards all output events until the time period passed as measured from the start of the statement, or until the number of output events are reached. The discarded events are not output and do not count towards any further output conditions if any are specified.

For example, the following statement outputs every minute the total price for all orders in the 30-minute time window but only after 30 minutes have passed:

select sum(price) from OrderEvent.win:time(30 min) output after 30 min snapshot every 1 min

An example in which after occur alone is below, in a statement that outputs total price for all orders in the last minute but only after 1 minute passed, each time an event arrives or leaves the data window:

select sum(price) from OrderEvent.win:time(1 min) output after 1 min

To demonstrate after when used with an event count, this statement find pairs of orders with the same id but suppresses output for the first 5 pairs:

select * from pattern[every o=OrderEvent->p=OrderEvent(id=o.id)] output after 5 events

4.7.2. Aggregation, Group By, Having and Output clause interaction

Remove stream events can also be useful in conjunction with aggregation and the output clause: When the engine posts remove stream events for fully-aggregated queries, it presents the aggregation state before the expiring event leaves the data window. Your application can thus easily obtain a delta between the new aggregation value and the prior aggregation value.

The engine evaluates the having-clause at the granularity of the data posted by views. That is, if you utilize a time window and output every 10 events, the having clause applies to each individual event or events entering and leaving the time window (and not once per batch of 10 events).

The output clause interacts in two ways with the group by and having clauses. First, in the output every n events case, the number n refers to the number of events arriving into the group by clause. That is, if the group by clause outputs only 1 event per group, or if the arriving events don't satisfy the having clause, then the actual number of events output by the statement could be fewer than n.

Second, the last, all and first keywords have special meanings when used in a statement with aggregate functions and the group by clause:

  • When no keyword is specified, the engine produces an output row for each row in the batch or when using group-by then an output per group only for those groups present in the batch, following Section 3.7.2, “Output for Aggregation and Group-By”.

  • The all keyword (the default) specifies that the most recent data for all groups seen so far should be output, whether or not these groups' aggregate values have just been updated

  • The last keyword specifies that only groups whose aggregate values have been updated with the most recent batch of events should be output.

  • The first keyword specifies that only groups whose aggregate values have been updated with the most recent batch of events should be output following the defined frequency, keeping frequency state for each group.

Please consult the Appendix A, Output Reference and Samples for detailed information on insert and remove stream output for aggregation and group-by.

By adding an output rate limiting clause to a statement that contains a group by clause we can control output of groups to obtain one row for each group, generating an event per group at the given output frequency.

The next statement outputs total price per symbol cumulatively (no data window was used here). As it specifies the all keyword, the statement outputs the current value for all groups seen so far, regardless of whether the group was updated in the last interval. Output occurs after an interval of 5 seconds passed and at the end of each subsequent interval:

select symbol, sum(price) from StockTickEvent group by symbol output all every 5 seconds

The below statement outputs total price per symbol considering events in the last 3 minutes. When events leave the 3-minute data window output also occurs as new aggregation values are computed. The last keyword instructs the engine to output only those groups that had changes. Output occurs after an interval of 10 seconds passed and at the end of each subsequent interval:

select symbol, sum(price) from StockTickEvent.win:time(3 min)
group by symbol output last every 10 seconds

This statement also outputs total price per symbol considering events in the last 3 minutes. The first keyword instructs the engine to output as soon as there is a new value for a group. After output for a given group the engine suppresses output for the same group for 10 seconds and does not suppress output for other groups. Output occurs again for that group after the interval when the group has new value(s):

select symbol, sum(price) from StockTickEvent.win:time(3 min)
group by symbol output first every 10 seconds

4.7.3. Runtime Considerations

Output rate limiting provides output events to your application in regular intervals. Between intervals, the engine uses a buffer to hold events until the output condition is reached. If your application has high-volume streams, you may need to be mindful of the memory needs for output rates.

The output clause with the snapshot keyword does not require a buffer, all other output keywords do consume memory until the output condition is reached.

4.8. Sorting Output: the Order By Clause

The order by clause is optional. It is used for ordering output events by their properties, or by expressions involving those properties. .

For example, the following statement outputs batches of 5 or more stock tick events that are sorted first by price ascending and then by volume ascending:

select symbol from StockTickEvent.win:time(60 sec) 
output every 5 events 
order by price, volume

Here is the syntax for the order by clause:

order by expression [asc | desc] [, expression [asc | desc]] [, ...]

If the order by clause is absent then the engine still makes certain guarantees about the ordering of output:

  • If the statement is not a join, does not group via group by clause and does not declare grouped data windows via std:groupwin view, the order in which events are delivered to listeners and through the iterator pull API is the order of event arrival.

  • If the statement is a join or outer join, or groups, then the order in which events are delivered to listeners and through the iterator pull API is not well-defined. Use the order by clause if your application requires events to be delivered in a well-defined order.

Esper places the following restrictions on the expressions in the order by clause:

  1. All aggregate functions that appear in the order by clause must also appear in the select expression.

Otherwise, any kind of expression that can appear in the select clause, as well as any name defined in the select clause, is also valid in the order by clause.

By default all sort operations on string values are performed via the Compare method and are thus not locale dependent. To account for differences in language or locale, see Section 11.4.17, “Engine Settings related to Language and Locale” to change this setting.

4.9. Limiting Row Count: the Limit Clause

The limit clause is typically used together with the order by and output clause to limit your query results to those that fall within a specified range. You can use it to receive the first given number of result rows, or to receive a range of result rows.

There are two syntaxes for the limit clause, each can be parameterized by integer constants or by variable names. The first syntax is shown below:

limit row_count [offset offset_count]

The required row_count parameter specifies the number of rows to output. The row_count can be an integer constant and can also be the name of the integer-type variable to evaluate at runtime.

The optional offset_count parameter specifies the number of rows that should be skipped (offset) at the beginning of the result set. A variable can also be used for this parameter.

The next sample EPL query outputs the top 10 counts per property 'uri' every 1 minute.

select uri, count(*) from WebEvent 
group by uri 
output snapshot every 1 minute
order by count(*) desc 
limit 10

The next statement demonstrates the use of the offset keyword. It outputs ranks 3 to 10 per property 'uri' every 1 minute:

select uri, count(*) from WebEvent 
group by uri 
output snapshot every 1 minute
order by count(*) desc 
limit 8 offset 2

The second syntax for the limit clause is for SQL standard compatibility and specifies the offset first, followed by the row count:

limit offset_count[, row_count]

The following are equivalent:

limit 8 offset 2
// ...equivalent to
limit 2, 8

A negative value for row_count returns an unlimited number or rows, and a zero value returns no rows. If variables are used, then the current variable value at the time of output dictates the row count and offset. A variable returning a null value for row_count also returns an unlimited number or rows.

A negative value for offset is not allowed. If your variable returns a negative or null value for offset then the value is assumed to be zero (i.e. no offset).

The iterator pull API also honors the limit clause, if present.

4.10. Merging Streams and Continuous Insertion: the Insert Into Clause

The insert into clause is optional in Esper. The clause can be specified to make the results of a statement available as an event stream for use in further statements, or to insert events into a named window. The clause can also be used to merge multiple event streams to form a single stream of events.

The syntax for the insert into clause is as follows:

insert [istream | rstream] into event_stream_name  [ (property_name [, property_name] ) ]

The istream (default) and rstream keywords are optional. If no keyword or the istream keyword is specified, the engine supplies the insert stream events generated by the statement. The insert stream consists of the events entering the respective window(s) or stream(s). If the rstream keyword is specified, the engine supplies the remove stream events generated by the statement. The remove stream consists of the events leaving the respective window(s).

The event_stream_name is an identifier that names the event stream (and also implicitly names the types of events in the stream) generated by the engine. The identifier can be used in further statements to filter and process events of that event stream. The insert into clause can consist of just an event stream name, or an event stream name and one or more property names.

The engine also allows listeners to be attached to a statement that contain an insert into clause. Listeners receive all events posted to the event stream.

To merge event streams, simply use the same event_stream_name identifier in all EPL statements that merge their result event streams. Make sure to use the same number and names of event properties and event property types match up.

Esper places the following restrictions on the insert into clause:

  1. The number of elements in the select clause must match the number of elements in the insert into clause if the clause specifies a list of event property names

  2. If the event stream name has already been defined by a prior statement or configuration, and the event property names and/or event types do not match, an exception is thrown at statement creation time.

The following sample inserts into an event stream by name CombinedEvent:

insert into CombinedEvent
select A.customerId as custId, A.timestamp - B.timestamp as latency
  from EventA.win:time(30 min) A, EventB.win:time(30 min) B
 where A.txnId = B.txnId

Each event in the CombinedEvent event stream has two event properties named "custId" and "latency". The events generated by the above statement can be used in further statements, such as shown in the next statement:

select custId, sum(latency)
  from CombinedEvent.win:time(30 min)
 group by custId

The example statement below shows the alternative form of the insert into clause that explicitly defines the property names to use.

insert into CombinedEvent (custId, latency)
select A.customerId, A.timestamp - B.timestamp 
...

The rstream keyword can be useful to indicate to the engine to generate only remove stream events. This can be useful if we want to trigger actions when events leave a window rather then when events enter a window. The statement below generates CombinedEvent events when EventA and EventB leave the window after 30 minutes.

insert rstream into CombinedEvent
select A.customerId as custId, A.timestamp - B.timestamp as latency
  from EventA.win:time(30 min) A, EventB.win:time(30 min) B
 where A.txnId = B.txnId

The insert into clause can be used in connection with patterns to provide pattern results to further statements for analysis:

insert into ReUpEvent
select linkUp.ip as ip 
from pattern [every linkDown=LinkDownEvent -> linkUp=LinkUpEvent(ip=linkDown.ip)]

4.10.1. Transposing a Property To a Stream

Sometimes your events may carry properties that are themselves event objects. Therefore EPL offers a special syntax to insert the value of a property itself as an event into a stream:

insert into stream_name select property_name.* from ...

This feature is only supported for native events and is not supported for Map or XML events. Nested property names are also not supported.

In this example, the class Summary with properties bid and ask that are of type Quote is:

public class Summary {
  private Quote bid;
  private Quote ask;
  ...

The statement to populate a stream of Quote events is thus:

insert into MyBidStream select bid.* from Summary

4.10.2. Merging Streams By Event Type

The insert into clause allows to merge multiple event streams into a event single stream. The clause names an event stream to insert into by specifing an event_stream_name. The first statement that inserts into the named stream defines the stream's event types. Further statements that insert into the same event stream must match the type of events inserted into the stream as declared by the first statement.

One approach to merging event streams specifies individual colum names either in the select clause or in the insert into clause of the statement. This approach has been shown in earlier examples.

Another approach to merging event streams specifies the wildcard (*) in the select clause (or the stream wildcard) to select the underlying event. The events in the event stream must then have the same event type as generated by the from clause.

Assume a statement creates an event stream named MergedStream by selecting OrderEvent events:

insert into MergedStream select * from OrderEvent

A statement can use the stream wildcard selector to select only OrderEvent events in a join:

insert into MergedStream select ord.* from ItemScanEvent, OrderEvent as ord

And a statement may also use an application-supplied user-defined function to convert events to OrderEvent instances:

insert into MergedStream select MyLib.convert(item) from ItemScanEvent as item

Esper specifically recognizes a conversion function: A conversion function must be the only selected column, and it must return either a CLR object or System.Collections.Generic.IDictionary.

4.10.3. Merging Disparate Types of Events: Variant Streams

A variant stream is a predefined stream into which events of multiple disparate event types can be inserted.

A variant stream name may appear anywhere in a pattern or from clause. In a pattern, a filter against a variant stream matches any events of any of the event types inserted into the variant stream. In a from clause including for named windows, views declared onto a variant stream may hold events of any of the event types inserted into the variant stream.

A variant stream is thus useful in problems that require different types of event to be treated the same.

Variant streams can be declared by means of create variant schema or can be predefined via runtime or initialization-time configuration as described in Section 11.4.23, “Variant Stream”. Your application may declare or predefine variant streams to carry events of a limited set of event types, or you may choose the variant stream to carry any and all types of events. This choice affects what event properties are available for consuming statements or patterns of the variant stream.

Assume that an application predefined a variant stream named OrderStream to carry only ServiceOrder and ProductOrder events. An insert into clause inserts events into the variant stream:

insert into OrderStream select * from ServiceOrder
insert into OrderStream select * from ProductOrder

Here is a sample statement that consumes the variant stream and outputs a total price per customer id for the last 30 seconds of ServiceOrder and ProductOrder events:

select customerId, sum(price) from OrderStream.win:time(30 sec) group by customerId

If your application predefines the variant stream to hold specific type of events, as the sample above did, then all event properties that are common to all specified types are visible on the variant stream, including nested, indexed and mapped properties. For access to properties that are only available on one of the types, the dynamic property syntax must be used. In the example above, the customerId and price were properties common to both ServiceOrder and ProductOrder events.

For example, here is a consuming statement that selects a service duraction property that only ServiceOrder events have, and that must therefore be casted to double and null values removed in order to aggregate:

select customerId, sum(coalesce(cast(serviceDuraction?, double), 0)) 
from OrderStream.win:time(30 sec) group by customerId

If your application predefines a variant stream to hold any type of events (the any type variance), then all event properties of the variant stream are effectively dynamic properties.

For example, an application may define an OutgoingEvents variant stream to hold any type of event. The next statement is a sample consumer of the OutgoingEvents variant stream that looks for the destination property and fires for each event in which the property exists with a value of 'email':

select * from OutgoingEvents(destination = 'email')

4.10.4. Decorated Events

Your select clause may use the '*' wildcard together with further expressions to populate a stream of events. A sample statement is:

insert into OrderStream select *, price*units as linePrice from PurchaseOrder

When using wildcard and selecting additional expression results, the engine produces what is called decorating events for the resulting stream. Decorating events add additional property values to an underlying event.

In the above example the resulting OrderStream consists of underlying PurchaseOrder events decorated by a linePrice property that is a result of the price*units expression.

In order to use insert into to insert into an existing stream of decorated events, your underlying event type must match, and all additional decorating property names and types of the select clause must also match.

4.10.5. Event as a Property

Your select clause may use the stream name to populate a stream of events in which each event has properties that are itself an event. A sample statement is:

insert into CompositeStream select order, service, order.price+service.price as totalPrice 
from PurchaseOrder.std:lastevent() as order, ServiceEvent:std:lastevent() as service

When using the stream name (or tag in patterns) in the select clause, the engine produces composite events: One or more of the properties of the composite event are events themselves.

In the above example the resulting CompositeStream consists of 3 columns: the PurchaseOrder event, the ServiceEvent event and the totalPrice property that is a result of the order.price+service.price expression.

In order to use insert into to insert into an existing stream of events in which properties are themselves events, each event column's event type must match, and all additional property names and types of the select clause must also match.

4.10.6. Populating an Underlying Event Object

Your insert into clause may also directly instantiate and populate application underlying event objects or Map event objects. This is described in greater detail in Section 2.11, “Event Objects Populated by Insert Into”.

4.11. Joining Event Streams

Two or more event streams can be part of the from clause and thus both (all) streams determine the resulting events. The where clause lists the join conditions that Esper uses to relate events in the two or more streams. Reference and historical data such as stored in your relational database, and data returned by a method invocation, can also be included in joins. Please see Section 4.15, “Accessing Relational Data via SQL” and Section 4.16, “Accessing Non-Relational Data via Method Invocation” for details.

Each point in time that an event arrives to one of the event streams, the two event streams are joined and output events are produced according to the where clause.

This example joins 2 event streams. The first event stream consists of fraud warning events for which we keep the last 30 minutes. The second stream is withdrawal events for which we consider the last 30 seconds. The streams are joined on account number.

select fraud.accountNumber as accntNum, fraud.warning as warn, withdraw.amount as amount,
       max(fraud.timestamp, withdraw.timestamp) as timestamp, 'withdrawlFraud' as desc
  from com.espertech.esper.example.atm.FraudWarningEvent.win:time(30 min) as fraud,
       com.espertech.esper.example.atm.WithdrawalEvent.win:time(30 sec) as withdraw
 where fraud.accountNumber = withdraw.accountNumber

Joins can also include one or more pattern statements as the next example shows:

select * from FraudWarningEvent.win:time(30 min) as fraud,
    pattern [every w=WithdrawalEvent -> PINChangeEvent(acct=w.acct)].std:lastevent() as withdraw
 where fraud.accountNumber = withdraw.w.accountNumber

The statement above joins the last 30 minutes of fraud warnings with a pattern. The pattern consists of every withdrawal event that is followed by a PIN change event for the same account number. It joins the two event streams on account number. The last-event view instucts the join to only consider the last pattern match.

In a join and outer join, your statement must declare a data window view or other view onto each stream. Streams that are marked as unidirectional and named windows as well as database or methods in a join are an exception and do not require a view to be specified. If you are joining an event to itself via contained-event selection, views also do not need to be specified.

The next example joins all FraudWarningEvent events that arrived since the statement was started, with the last 20 seconds of PINChangeEvent events:

select * from FraudWarningEvent.win:keepall() as fraud, PINChangeEvent.win:time(20 sec) as pin
 where fraud.accountNumber = pin.accountNumber

The above example employed the special keep-all view that retains all events.

4.12. Outer and Inner Joins

Esper supports left outer joins, right outer joins, full outer joins and inner joins in any combination between an unlimited number of event streams. Outer and inner joins can also join reference and historical data as explained in Section 4.15, “Accessing Relational Data via SQL”, as well as join data returned by a method invocation as outlined in Section 4.16, “Accessing Non-Relational Data via Method Invocation”.

The keywords left, right, full and inner control the type of the join between two streams. The on clause specifies one or more properties that join each stream. The synopsis is as follows:

...from stream_def [as name] 
  ((left|right|full outer) | inner) join stream_def 
  on property = property [and property = property ...]
  [ ((left|right|full outer) | inner) join stream_def on ...]...

If the outer join is a left outer join, there will be an output event for each event of the stream on the left-hand side of the clause. For example, in the left outer join shown below we will get output for each event in the stream RfidEvent, even if the event does not match any event in the event stream OrderList.

select * from RfidEvent.win:time(30 sec) as rfid
       left outer join
       OrderList.win:length(10000) as orderlist
     on rfid.itemId = orderList.itemId

Similarly, if the join is a Right Outer Join, then there will be an output event for each event of the stream on the right-hand side of the clause. For example, in the right outer join shown below we will get output for each event in the stream OrderList, even if the event does not match any event in the event stream RfidEvent.

select * from RfidEvent.win:time(30 sec) as rfid
       right outer join
       OrderList.win:length(10000) as orderlist
       on rfid.itemId = orderList.itemId

For all types of outer joins, if the join condition is not met, the select list is computed with the event properties of the arrived event while all other event properties are considered to be null.

The next type of outer join is a full outer join. In a full outer join, each point in time that an event arrives to one of the event streams, one or more output events are produced. In the example below, when either an RfidEvent or an OrderList event arrive, one or more output event is produced. The next example shows a full outer join that joins on multiple properties:

select * from RfidEvent.win:time(30 sec) as rfid
       full outer join
       OrderList.win:length(10000) as orderlist
       on rfid.itemId = orderList.itemId and rfid.assetId = orderList.assetId

The last type of join is an inner join. In an inner join, the engine produces an output event for each event of the stream on the left-hand side that matches at least one event on the right hand side considering the join properties. For example, in the inner join shown below we will get output for each event in the RfidEvent stream that matches one or more events in the OrderList data window:

select * from RfidEvent.win:time(30 sec) as rfid
       inner join
       OrderList.win:length(10000) as orderlist
       on rfid.itemId = orderList.itemId and rfid.assetId = orderList.assetId

Patterns as streams in a join follow this rule: If no data window view is declared for the pattern then the pattern stream retains the last match. Thus a pattern must have matched at least once for the last row to become available in a join. Multiple rows from a pattern stream may be retained by declaring a data window view onto a pattern using the pattern [...].view_specification syntax.

Finally, this example outer joins multiple streams. Here the RfidEvent stream is outer joined to both ProductName and LocationDescription via left outer join:

select * from RfidEvent.win:time(30 sec) as rfid
      left outer join ProductName.win:keepall() as refprod
        on rfid.productId = refprod.prodId
      left outer join LocationDescription.win:keepall() as refdesc
        on rfid.location = refdesc.locId

The on clause may only employ the = equals operator and property names. Any other operators must be placed in the where-clause.

4.13. Unidirectional Joins

In a join or outer join your statement lists multiple event streams, views and/or patterns in the from clause. As events arrive into the engine, each of the streams (views, patterns) provides insert and remove stream events. The engine evaluates each insert and remove stream event provided by each stream, and joins or outer joins each event against data window contents of each stream, and thus generates insert and remove stream join results.

The direction of the join execution depends on which stream or streams are currently providing an insert or remove stream event for executing the join. A join is thus multidirectional, or bidirectional when only two streams are joined. A join can be made unidirectional if your application does not want new results when events arrive on a given stream or streams.

The unidirectional keyword can be used in the from clause to identify a single stream that provides the events to execute the join. If the keyword is present for a stream, all other streams in the from clause become passive streams. When events arrive or leave a data window of a passive stream then the join does not generate join results.

For example, consider a use case that requires us to join stock tick events (TickEvent) and news events (NewsEvent). The unidirectional keyword allows to generate results only when TickEvent events arrive, and not when NewsEvent arrive or leave the 10-second time window:

select * from TickEvent unidirectional, NewsEvent.win:time(10 sec) 
where tick.symbol = news.symbol

Aggregation functions in a unidirectional join aggregate within the context of each unidirectional event evaluation and are not cumulative.

The count function in the next query returns, for each TickEvent, the number of matching NewEvent events:

select count(*) from TickEvent unidirectional, NewsEvent.win:time(10 sec) 
where tick.symbol = news.symbol

The following restrictions apply to unidirectional joins:

  1. The unidirectional keyword can only be specified for a single stream in the from clause.

  2. Receiving data from a unidirectional join via the pull API (GetEnumerator method) is not allowed. This is because the engine holds no state for the single stream that provides the events to execute the join.

  3. The stream that declares the unidirectional keyword cannot declare a data window view or other view for that stream, since remove stream events are not processed for the single stream.

4.14. Subqueries

A subquery is a select within another statement. Esper supports subqueries in the select clause, in the where clause and in stream and pattern filter expressions. Subqueries provide an alternative way to perform operations that would otherwise require complex joins. Subqueries can also make statements more readable then complex joins.

Esper supports both simple subqueries as well as correlated subqueries. In a simple subquery, the inner query is not correlated to the outer query. Here is an example simple subquery within a select clause:

select assetId, (select zone from ZoneClosed.std:lastevent()) as lastClosed from RFIDEvent

If the inner query is dependent on the outer query, we will have a correlated subquery. An example of a correlated subquery is shown below. Notice the where clause in the inner query, where the condition involves a stream from the outer query:

select * from RfidEvent as RFID where 'Dock 1' = 
  (select name from Zones.std:unique(zoneId) where zoneId = RFID.zoneId)

The example above shows a subquery in the where clause. The statement selects RFID events in which the zone name matches a string constant based on zone id. The statement uses the view std:unique to guarantee that only the last event per zone id is held from processing by the subquery.

The next example is a correlated subquery within a select clause. In this statement the select clause retrieves the zone name by means of a subquery against the Zones set of events correlated by zone id:

select zoneId, (select name from Zones.std:unique(zoneId) 
  where zoneId = RFID.zoneId) as name from RFIDEvent

Note that when a simple or correlated subquery returns multiple rows, the engine returns a null value as the subquery result. To limit the number of events returned by a subquery consider using one of the views std:lastevent, std:unique and std:groupwin or aggregation functions or the multi-row and multi-column selects as described below.

The select clause of a subquery also allows wildcard selects, which return as an event property the underlying event object of the event type as defined in the from clause. An example:

select (select * from MarketData.std:lastevent()) as md 
  from pattern [every timer:interval(10 sec)]

The output events to the statement above contain the underlying MarketData event in a property named "md". The statement populates the last MarketData event into a property named "md" every 10 seconds following the pattern definition, or populates a null value if no MarketData event has been encountered so far.

When your subquery returns multiple rows, you must use an aggregation function in the select clause of the subselect, as a subquery can only return a single row and single value object. To return multiple values from a subquery, consider writing a custom aggregation function that returns an array or collection of values.

Aggregation functions may be used in the select clause of the subselect as this example outlines:

select * from MarketData
where price > (select max(price) from MarketData(symbol='GOOG').std:lastevent())

As the sub-select expression is evaluated first (by default), the query above actually never fires for the GOOG symbol, only for other symbols that have a price higher then the current maximum for GOOG. As a sidenote, the insert into clause can also be handy to compute aggregation results for use in multiple subqueries.

When using aggregation functions in a correlated subselect the engine computes the aggregation based on data window or named window contents matching the where-clause.

The following example compares the quantity value provided by the current order event against the total quantity of all order events in the last 1 hour for the same client.

select * from OrderEvent oe
where qty > 
  (select sum(qty) from OrderEvent.win:time(1 hour) pd 
  where pd.client = oe.client)

Filter expressions in a pattern or stream may also employ subqueries. Subqueries can be uncorrelated or can be correlated to properties of the stream or to properties of tagged events in a pattern. Subqueries may reference named windows as well.

The following example filters BarData events that have a close price less then the last moving average (field movAgv) as provided by stream SMA20Stream (an uncorrelated subquery):

select * from BarData(ticker='MSFT', closePrice < 
    (select movAgv from SMA20Stream(ticker='MSFT').std:lastevent()))

A few generic examples follow to demonstrate the point. The examples use short event and property names so they are easy to read. Assume A and B are streams and DNamedWindow is a named window, and properties a_id, b_id, d_id, a_val, b_val, d_val respectively:

// Sample correlated subquery as part of stream filter criteria
select * from A(a_val in 
  (select b_val from B.std:unique(b_val) as b where a.a_id = b.b_id)) as a
// Sample correlated subquery against a named window
select * from A(a_val in 
  (select b_val from DNamedWindow as d where a.a_id = d.d_id)) as a
// Sample correlated subquery in the filter criteria as part of a pattern, querying a named window
select * from pattern [
  a=A -> b=B(bvalue = 
    (select d_val from DNamedWindow as d where d.d_id = b.b_id and d.d_id = a.a_id))
]

Subquery state starts to accumulate as soon as a statement starts (and not only when a pattern-subexpression activates).

The following restrictions apply to subqueries:

  1. The subquery stream definition must define a data window or other view to limit subquery results, reducing the number of events held for subquery execution

  2. Subqueries can only consist of a select clause, a from clause and a where clause. The group by and having clauses, as well as joins, outer-joins and output rate limiting are not permitted within subqueries.

  3. If using aggregation functions in a subquery, note these limitations:

    1. None of the properties of the correlated stream(s) can be used within aggregation functions.

    2. The properties of the subselect stream must all be within aggregation functions.

The order of evaluation of subqueries relative to the containing statement is guaranteed: If the containing statement and its subqueries are reacting to the same type of event, the subquery will receive the event first before the containing statement's clauses are evaluated. This behavior can be changed via configuration. The order of evaluation of subqueries is not guaranteed between subqueries.

Performance of your statement containing one or more subqueries principally depends on two parameters. First, if your subquery correlates one or more columns in the subquery stream with the enclosing statement's streams via equals '=', the engine automatically builds the appropriate indexes for fast row retrieval based on the key values correlated (joined). The second parameter is the number of rows found in the subquery stream and the complexity of the filter criteria (where clause), as each row in the subquery stream must evaluate against the where clause filter.

4.14.1. The 'exists' Keyword

The exists condition is considered "to be met" if the subquery returns at least one row. The not exists condition is considered true if the subquery returns no rows.

The synopsis for the exists keyword is as follows:

exists (subquery)

Let's take a look at a simple example. The following is an EPL statement that uses the exists condition:

select assetId from RFIDEvent as RFID 
  where exists (select * from Asset.std:unique(assetId) where assetId = RFID.assetId)

This select statement will return all RFID events where there is at least one event in Assets unique by asset id with the same asset id.

4.14.2. The 'in' and 'not in' Keywords

The in subquery condition is true if the value of an expression matches one or more of the values returned by the subquery. Consequently, the not in condition is true if the value of an expression matches none of the values returned by the subquery.

The synopsis for the in keyword is as follows:

expression in (subquery)

The right-hand side subquery must return exactly one column.

The next statement demonstrates the use of the in subquery condition:

select assetId from RFIDEvent
  where zone in (select zone from ZoneUpdate(status = 'closed').win:time(10 min))

The above statement demonstrated the in subquery to select RFID events for which the zone status is in a closed state.

Note that if the left-hand expression yields null, or if there are no equal right-hand values and at least one right-hand row yields null, the result of the in construct will be null, not false (or true for not-in). This is in accordance with SQL's normal rules for Boolean combinations of null values.

4.14.3. The 'any' and 'some' Keywords

The any subquery condition is true if the expression returns true for one or more of the values returned by the subquery.

The synopsis for the any keyword is as follows:

expression operator any (subquery)
expression operator some (subquery)

The left-hand expression is evaluated and compared to each row of the subquery result using the given operator, which must yield a Boolean result. The result of any is "true" if any true result is obtained. The result is "false" if no true result is found (including the special case where the subquery returns no rows).

The operator can be any of the following values: =, !=, <>, <, <=, >, >=.

The some keyword is a synonym for any. The in construct is equivalent to = any.

The right-hand side subquery must return exactly one column.

The next statement demonstrates the use of the any subquery condition:

select * from ProductOrder as ord
  where quantity < any
    (select minimumQuantity from MinimumQuantity.win:keepall())

The above query compares ProductOrder event's quantity value with all rows from the MinimumQuantity stream of events and returns only those ProductOrder events that have a quantity that is less then any of the minimum quantity values of the MinimumQuantity events.

Note that if there are no successes and at least one right-hand row yields null for the operator's result, the result of the any construct will be null, not false. This is in accordance with SQL's normal rules for Boolean combinations of null values.

4.14.4. The 'all' Keyword

The all subquery condition is true if the expression returns true for all of the values returned by the subquery.

The synopsis for the all keyword is as follows:

expression operator all (subquery)

The left-hand expression is evaluated and compared to each row of the subquery result using the given operator, which must yield a Boolean result. The result of all is "true" if all rows yield true (including the special case where the subquery returns no rows). The result is "false" if any false result is found. The result is null if the comparison does not return false for any row, and it returns null for at least one row.

The operator can be any of the following values: =, !=, <>, <, <=, >, >=.

The not in construct is equivalent to != all.

The right-hand side subquery must return exactly one column.

The next statement demonstrates the use of the all subquery condition:

select * from ProductOrder as ord
  where quantity < all
    (select minimumQuantity from MinimumQuantity.win:keepall())

The above query compares ProductOrder event's quantity value with all rows from the MinimumQuantity stream of events and returns only those ProductOrder events that have a quantity that is less then all of the minimum quantity values of the MinimumQuantity events.

4.14.5. Multi-Column Selection

Your subquery may select multiple columns in the select clause including multiple aggregated values from a data window or named window.

The following example is a correlated subquery that selects wildcard and in addition selects the bid and offer properties of the last MarketData event for the same symbol as the arriving OrderEvent:

select *,
  (select bid, offer from MarketData.std:unique(symbol) as md 
   where md.symbol = oe.symbol) as bidoffer
from OrderEvent oe

Output events for the above query contain all properties of the original OrderEvent event. In addition each output event contains a bidoffer nested property that itself contains the bid and offer properties. You may retrieve the bid and offer from output events directly via the bidoffer.bid property name syntax for nested properties.

The next example is similar to the above query but instead selects aggregations and selects from a named window by name OrderNamedWindow (creation not shown here). For each arriving OrderEvent it selects the total quantity and count of all order events for the same client, as currently held by the named window:

select *,
  (select sum(qty) as sumPrice, count(*) as countRows 
   from OrderNamedWindow as onw
   where onw.client = oe.client) as pastOrderTotals
from OrderEvent as oe

Output events for the above query contain all properties of the original OrderEvent event. In addition each output event contains a pastOrderTotals nested property that itself contains the sumPrice and countRows properties.

4.14.6. Multi-Row Selection

While a subquery cannot change the cardinality of the selected stream, a subquery can return multiple values from the selected data window or named window. This section shows examples of the window aggregation function in subselects.

Consider using an inner join, outer join or unidirectional join instead to achieve a 1-to-many cardinality in the number of output events.

The next example is an uncorrelated subquery that selects all current ZoneEvent events considering the last ZoneEvent per zone for each arriving RFIDEvent.

select assetId,
 (select window(z.*) as winzones from ZoneEvent.std:unique(zone) as z) as zones
 from RFIDEvent

Output events for the above query contain two properties: the assetId property and the zones property. The latter property is a nested property that contains the winzones property. You may retrieve the zones from output events directly via the zones.winzones property name syntax for nested properties.

In this example for a correlated subquery against a named window we assume that the OrderNamedWindow has been created and contains order events. The query returns for each MarketData event the list of order ids for orders with the same symbol:

select price,
 (select window(orderId) as winorders 
  from OrderNamedWindow onw 
  where onw.symbol = md.symbol) as orderIds
 from MarketData md

Output events for the above query contain two properties: the price property and the orderIds property. The latter property is a nested property that contains the winorders property of type array.

4.14.7. Hints Related to Subqueries

The following hints are available to tune performance and memory use of subqueries.

Use the @Hint('set_noindex') hint for a statement that utilizes one or more subqueries. It instructs the engine to always perform a full table scan. The engine does not build an implicit index or use an explicitly-created index when this hint is provided. Use of the hint may result in reduced memory use but poor statement performance.

The following hints are available to tune performance and memory use of subqueries that select from named windows.

Named windows are globally-visible data windows. As such an application may create explicit indexes as discussed in Section 4.17.10, “Explicitly Indexing Named Windows”. The engine may also elect to create implicit indexes (no create-index EPL required) for index-based lookup of rows when executing on-select, on-merge, on-update and on-delete statements and for statements that subquery a named window.

By default and without specifying a hint, each statement that subqueries a named window also maintains its own index for looking up events held by the named window. The engine maintains the index by consuming the named window insert and remove stream. When the statement is destroyed it releases that index.

Specify the @Hint('enable_window_subquery_indexshare') hint to enable subquery index sharing for named windows. When using this hint, indexes for subqueries are maintained by the named window itself (and not each statement), are shared between one or more statements and may also utilize explicit indexes. Specify the hint once as part of the create window statement.

This sample EPL statement creates a named window with subquery index sharing enabled:

@Hint('enable_window_subquery_indexshare')
create window AllOrdersNamedWindow.win:keepall() as OrderMapEventType

When subquery index sharing is enabled, performance may increase as named window stream consumption is no longer needed. You may also expect reduced memory use especially if a large number of EPL statements perform similar subqueries against a named window. Subquery index sharing may require additional short-lived object creation and may slightly increase lock held time for named windows.

The following statement performs a correlated subquery against the named window above. When a settlement event arrives it select the order detail for the same order id as provided by the settlement event:

select 
  (select * from AllOrdersNamedWindow as onw 
    where onw.orderId = se.orderId) as orderDetail
  from SettlementEvent as se

With subquery index sharing enabled the engine maintains an index of order events by order id for the named window, and shares that index between additional statements until the time all utilizing statements are destroyed.

You may disable subquery index sharing for a specific statement by specifying the @Hint('disable_window_subquery_indexshare') hint, as this example shows, causing the statement to maintain its own index:

@Hint('disable_window_subquery_indexshare')
select 
  (select * from AllOrdersNamedWindow as onw 
    where onw.orderId = se.orderId) as orderDetail
  from SettlementEvent as se

4.15. Accessing Relational Data via SQL

This chapter outlines how reference data and historical data that are stored in a relational database can be queried via SQL within EPL statements.

Esper can access via join and outer join as well as via iterator (poll) API all types of event streams to stored data. In order for such data sources to become accessible to Esper, some configuration is required. The Section 11.4.8, “Relational Database Access” explains the required configuration for database access in greater detail, and includes information on configuring a query result cache.

Esper does not parse or otherwise inspect your SQL query. Therefore your SQL can make use of any database-specific SQL language extensions or features that your database provides.

If you have enabled query result caching in your Esper database configuration, Esper retains SQL query results in cache following the configured cache eviction policy.

Also if you have enabled query result caching in your Esper database configuration and provide EPL where clause and/or on clause (outer join) expressions, then Esper builds indexes on the SQL query results to enable fast lookup. This is especially useful if your queries return a large number of rows. For building the proper indexes, Esper inspects the expression found in your EPL query where clause, if present. For outer joins, Esper also inspects your EPL query on clause. Esper analyzes the EPL on clause and where clause expressions, if present, looking for property comparison with or without logical AND-relationships between properties. When a SQL query returns rows for caching, Esper builds and caches the appropriate index and lookup strategies for fast row matching against indexes.

Joins or outer joins in which only SQL statements or method invocations are listed in the from clause and no other event streams are termed passive joins. A passive join does not produce an insert or remove stream and therefore does not invoke statement listeners with results. A passive join can be iterated on (polled) using a statement's GetSafeEnumerator and GetEnumerator methods.

There are no restrictions to the number of SQL statements or types of streams joined. The following restrictions currently apply:

  • Sub-views on an SQL query are not allowed; That is, one cannot create a time or length window on an SQL query. However one can use the insert into syntax to make join results available to a further statement.

  • Your database software must support JDBC prepared statements that provide statement meta data at compilation time. Most major databases provide this function. A workaround is available for databases that do not provide this function.

  • JDBC drivers must support the getMetadata feature. A workaround is available as below for JDBC drivers that don't support getting metadata.

The next sections assume basic knowledge of SQL (Structured Query Language).

4.15.1. Joining SQL Query Results

To join an event stream against stored data, specify the sql keyword followed by the name of the database and a parameterized SQL query. The syntax to use in the from clause of an EPL statement is:

sql:database_name [" parameterized_sql_query "]

The engine uses the database_name identifier to obtain configuration information in order to establish a database connection, as well as settings that control connection creation and removal. Please see Section 11.4.8, “Relational Database Access” to configure an engine for database access.

Following the database name is the SQL query to execute. The SQL query can contain one or more substitution parameters. The SQL query string is placed in single brackets [ and ]. The SQL query can be placed in either single quotes (') or double quotes ("). The SQL query grammer is passed to your database software unchanged, allowing you to write any SQL query syntax that your database understands, including stored procedure calls.

Substitution parameters in the SQL query string take the form ${expression}. The engine resolves expression at statement execution time to the actual expression result by evaluating the events in the joined event stream or current variable values, if any event property references or variables occur in the expression. An expression may not contain EPL substitution parameters.

The engine determines the type of the SQL query output columns by means of the result set metadata that your database software returns for the statement. The actual query results are obtained via System.Data.DbDataReader.

The sample EPL statement below joins an event stream consisting of CustomerCallEvent events with the results of an SQL query against the database named MyCustomerDB and table Customer:

select custId, cust_name from CustomerCallEvent,
  sql:MyCustomerDB [' select cust_name from Customer where cust_id = ${custId} ']

The example above assumes that CustomerCallEvent supplies an event property named custId. The SQL query selects the customer name from the Customer table. The where clause in the SQL matches the Customer table column cust_id with the value of custId in each CustomerCallEvent event. The engine executes the SQL query for each new CustomerCallEvent encountered.

If the SQL query returns no rows for a given customer id, the engine generates no output event. Else the engine generates one output event for each row returned by the SQL query. An outer join as described in the next section can be used to control whether the engine should generate output events even when the SQL query returns no rows.

The next example adds a time window of 30 seconds to the event stream CustomerCallEvent. It also renames the selected properties to customerName and customerId to demonstrate how the naming of columns in an SQL query can be used in the select clause in the EPL query. And the example uses explicit stream names via the as keyword.

select customerId, customerName from
  CustomerCallEvent.win:time(30 sec) as cce,
  sql:MyCustomerDB ["select cust_id as customerId, cust_name as customerName from Customer 
                  where cust_id = ${cce.custId}"] as cq

Any window, such as the time window, generates insert stream (istream) events as events enter the window, and remove stream (rstream) events as events leave the window. The engine executes the given SQL query for each CustomerCallEvent in both the insert stream and the remove stream. As a performance optimization, the istream or rstream keywords in the select clause can be used to instruct the engine to only join insert stream or remove stream events, reducing the number of SQL query executions.

Since any expression may be placed within the ${...} syntax, you may use variables or user-defined functions as well.

The next example assumes that a variable by name varLowerLimit is defined and that a user-defined function getLimit exists on the MyLib imported class that takes a LimitEvent as a parameter:

select * from LimitEvent le, 
  sql:MyCustomerDB [' select cust_name from Customer where 
      amount > ${max(varLowerLimit, MyLib.getLimit(le))} ']

The example above takes the higher of the current variable value or the value returned by the user-defined function to return only those customer names where the amount exceeds the computed limit.

4.15.2. SQL Query and the EPL Where Clause

Consider using the EPL where clause to join the SQL query result to your event stream. Similar to EPL joins and outer-joins that join event streams or patterns, the EPL where clause provides join criteria between the SQL query results and the event stream (as a side note, an SQL where clause is a filter of rows executed by your database on your database server before returning SQL query results).

Esper analyzes the expression in the EPL where clause and outer-join on clause, if present, and builds the appropriate indexes from that information at runtime, to ensure fast matching of event stream events to SQL query results, even if your SQL query returns a large number of rows. Your applications must ensure to configure a cache for your database using Esper configuration, as such indexes are held with regular data in a cache. If you application does not enable caching of SQL query results, the engine does not build indexes on cached data.

The sample EPL statement below joins an event stream consisting of OrderEvent events with the results of an SQL query against the database named MyRefDB and table SymbolReference:

select symbol, symbolDesc from OrderEvent as orders,
  sql:MyRefDB ['select symbolDesc from SymbolReference'] as reference
  where reference.symbol = orders.symbol

Notice how the EPL where clause joins the OrderEvent stream to the SymbolReference table. In this example, the SQL query itself does not have a SQL where clause and therefore returns all rows from table SymbolReference.

If your application enables caching, the SQL query fires only at the arrival of the first OrderEvent event. When the second OrderEvent arrives, the join execution uses the cached query result. If the caching policy that you specified in the Esper database configuration evicts the SQL query result from cache, then the engine fires the SQL query again to obtain a new result and places the result in cache.

If SQL result caching is enabled and your EPL where clause, as show in the above example, provides the properties to join, then the engine indexes the SQL query results in cache and retains the index together with the query result in cache. Thus your application can benefit from high performance index-based lookups as long as the SQL query results are found in cache.

The SQL result caches operate on the level of all result rows for a given parameter set. For example, if your query returns 10 rows for a certain set of parameter values then the cache treats all 10 rows as a single entry keyed by the parameter values, and the expiry policy applies to all 10 rows and not to each individual row.

It is also possible to join multiple autonomous database systems in a single query, for example:

select symbol, symbolDesc from OrderEvent as orders,
  sql:My_Oracle_DB ['select symbolDesc from SymbolReference'] as reference,
  sql:My_MySQL_DB ['select orderList from orderHistory'] as history
  where reference.symbol = orders.symbol
  and history.symbol = orders.symbol 

4.15.3. Outer Joins With SQL Queries

You can use outer joins to join data obtained from an SQL query and control when an event is produced. Use a left outer join, such as in the next statement, if you need an output event for each event regardless of whether or not the SQL query returns rows. If the SQL query returns no rows, the join result populates null values into the selected properties.

select custId, custName from
  CustomerCallEvent as cce
  left outer join 
  sql:MyCustomerDB ["select cust_id, cust_name as custName 
                     from Customer where cust_id = ${cce.custId}"] as cq
  on cce.custId = cq.cust_id

The statement above always generates at least one output event for each CustomerCallEvent, containing all columns selected by the SQL query, even if the SQL query does not return any rows. Note the on expression that is required for outer joins. The on acts as an additional filter to rows returned by the SQL query.

4.15.4. Using Patterns to Request (Poll) Data

Pattern statements and SQL queries can also be applied together in useful ways. One such use is to poll or request data from a database at regular intervals or following the schedule of the crontab-like timer:at.

The next statement is an example that shows a pattern that fires every 5 seconds to query the NewOrder table for new orders:

insert into NewOrders
select orderId, orderAmount from
  pattern [every timer:interval(5 sec)],
  sql:MyCustomerDB ['select orderId, orderAmount from NewOrders']

4.15.5. Polling SQL Queries via Iterator

Usually your SQL query will take part in a join and thus be triggered by an event or pattern occurrence. Instead, your application may need to poll a SQL query and thus use Esper query execution and caching facilities and obtain event data and metadata.

Your EPL statement can specify an SQL statement without a join. Such a stand-alone SQL statement does not post new events, and may only be queried via the iterator poll API. Your EPL and SQL statement may still use variables.

The next statement assumes that a price_var variable has been declared. It selects from the relational database table named NewOrder all rows in which the price column is greater then the current value of the price_var EPL variable:

select * from sql:MyCustomerDB ['select * from NewOrder where ${price_var} > price']

Use the iterator and GetSafeEnumerator methods on EPStatement to obtain results. The statement does not post events to listeners, it is strictly passive in that sense.

4.15.6. ADO.NET Implementation Overview

The engine translates SQL queries into JDBC System.Data.DbCommand statements by replacing ${name} parameters with '?' placeholders. It obtains name and type of result columns from the compiled PreparedStatement meta data when the EPL statement is created.

The engine supplies parameters to the compiled statement via the SetParameter method on DbCommand. The engine uses the DbDataReader to obtain column values.

4.15.7. No-Metadata Workaround

Certain database drivers are known to not return metadata for precompiled prepared SQL statements. This can be a problem as metadata is required by Esper. Esper obtains SQL result set metadata to validate an EPL statement and to provide column types for output events. Drivers that do not provide metadata for precompiled SQL statements require a workaround. Such drivers do generally provide metadata for executed SQL statements, however do not provide the metadata for precompiled SQL statements.

Please consult the Chapter 11, Configuration for the configuration options available in relation to metadata retrieval.

To obtain metadata for an SQL statement, Esper can alternatively fire a SQL statement which returns the same column names and types as the actual SQL statement but without returning any rows. This kind of SQL statement is referred to as a sample statement in below workaround description. The engine can then use the sample SQL statement to retrieve metadata for the column names and types returned by the actual SQL statement.

Applications can provide a sample SQL statement to retrieve metadata via the metadatasql keyword:

sql:database_name ["parameterized_sql_query" metadatasql "sql_meta_query"] 

The sql_meta_query must be an SQL statement that returns the same number of columns, the same type of columns and the same column names as the parameterized_sql_query, and does not return any rows.

Alternatively, applications can choose not to provide an explicit sample SQL statement. If the EPL statement does not use the metadatasql syntax, the engine applies lexical analysis to the SQL statement. From the lexical analysis Esper generates a sample SQL statement adding a restrictive clause "where 1=0" to the SQL statement.

Alternatively, you can add the following tag to the SQL statement: ${$ESPER-SAMPLE-WHERE}. If the tag exists in the SQL statement, the engine does not perform lexical analysis and simply replaces the tag with the SQL where clause "where 1=0". Therefore this workaround is applicable to SQL statements that cannot be correctly lexically analyzed. The SQL text after the placeholder is not part of the sample query. For example:

select mycol from sql:myDB [
  'select mycol from mytesttable ${$ESPER-SAMPLE-WHERE} where ....'], ...

If your parameterized_sql_query SQL query contains vendor-specific SQL syntax, generation of the metadata query may fail to produce a valid SQL statement. If you experience an SQL error while fetching metadata, use any of the above workarounds with the Oracle JDBC driver.

4.15.8. SQL Input Parameter and Column Output Conversion

As part of database access configuration you may optionally specify SQL type mappings. These mappings apply to all queries against the same database identified by name.

If your application must perform SQL-query-specific or EPL-statement-specific mapping or conversion between types, the facility to register a conversion callback exists as follows.

Use the @Hook instruction and HookType.SQLCOL as part of your EPL statement text to register a statement SQL parameter or column conversion hook. Implement the interface com.espertech.esper.client.hook.SQLColumnTypeConversion to perform the input parameter or column value conversion.

A sample statement with annotation is shown:

@Hook(type=HookType.SQLCOL, hook='MyDBTypeConvertor')
select * from sql:MyDB ['select * from MyEventTable]

The engine expects MyDBTypeConvertor to resolve to a class (considering engine imports) and instantiates one instance of MyDBTypeConvertor for each statement.

4.15.9. SQL Row Object Conversion

Your application may also directly convert a SQL result row into a type which is an opportunity for your application to interrogate and transform the SQL row result data freely before packing the data into a type. Your application can additionally indicate to skip SQL result rows.

Use the @Hook instruction and HookType.SQLROW as part of your EPL statement text to register a statement SQL output row conversion hook. Implement the interface com.espertech.esper.client.hook.SQLOutputRowConversion to perform the output row conversion.

A sample statement with annotation is shown:

@Hook(type=HookType.SQLROW, hook='MyDBRowConvertor')
select * from sql:MyDB ['select * from MyEventTable]

The engine expects MyDBRowConvertor to resolve to a class (considering engine imports) and instantiates one MyDBRowConvertor instance for each statement.

4.16. Accessing Non-Relational Data via Method Invocation

Your application may need to join data that originates from a web service, a distributed cache, an object-oriented database or simply data held in memory by your application. Esper accommodates this need by allowing a method invocation (or procedure call or function) in the from clause of a statement.

The results of such a method invocation in the from clause plays the same role as a relational database table in an inner and outer join in SQL. The role is thus dissimilar to the role of a user-defined function, which may occur in any expression such as in the select clause or the where clause. Both are backed by one or more static methods provided by your class library.

Esper can join and outer join an unlimited number and all types of event streams to the data returned by your method invocation. In addition, Esper can be configured to cache the data returned by your method invocations.

Joins or outer joins in which only SQL statements or method invocations are listed in the from clause and no other event streams are termed passive joins. A passive join does not produce an insert or remove stream and therefore does not invoke statement listeners with results. A passive join can be iterated on (polled) using a statement's GetSafeEnumerator and GetEnumerator methods.

The following restrictions currently apply:

  • Sub-views on a method invocations are not allowed; That is, one cannot create a time or length window on a method invocation. However one can use the insert into syntax to make join results available to a further statement.

4.16.1. Joining Method Invocation Results

The syntax for a method invocation in the from clause of an EPL statement is:

method:class_name.method_name[(parameter_expressions)]

The method keyword denotes a method invocation. It is followed by a class name and a method name separated by a dot (.) character. If you have parameters to your method invocation, these are placed in round brackets after the method name. Any expression is allowed as a parameter, and individual parameter expressions are separated by a comma. Expressions may also use event properties of the joined stream.

In the sample join statement shown next, the method 'lookupAsset' provided by class 'MyLookupLib' returns one or more rows based on the asset id (a property of the AssetMoveEvent) that is passed to the method:

select * from AssetMoveEvent, method:MyLookupLib.lookupAsset(assetId)

The following statement demonstrates the use of the where clause to join events to the rows returned by a method invocation, which in this example does not take parameters:

select assetId, assetDesc from AssetMoveEvent as asset, 
       method:MyLookupLib.getAssetDescriptions() as desc 
where asset.assetid = desc.assetid

Your method invocation may return zero, one or many rows for each method invocation. If you have caching enabled through configuration, then Esper can avoid the method invocation and instead use cached results. Similar to SQL joins, Esper also indexes cached result rows such that join operations based on the where clause or outer-join on clause can be very efficient, especially if your method invocation returns a large number of rows.

If the time taken by method invocations is critical to your application, you may configure local caches as Section 11.4.6, “Cache Settings for From-Clause Method Invocations” describes.

Esper analyzes the expression in the EPL where clause and outer-join on clause, if present, and builds the appropriate indexes from that information at runtime, to ensure fast matching of event stream events to method invocation results, even if your method invocation returns a large number of rows. Your applications must ensure to configure a cache for your method invocation using Esper configuration, as such indexes are held with regular data in a cache. If you application does not enable caching of method invocation results, the engine does not build indexes on cached data.

4.16.2. Polling Method Invocation Results via Iterator

Usually your method invocation will take part in a join and thus be triggered by an event or pattern occurrence. Instead, your application may need to poll a method invocation and thus use Esper query execution and caching facilities and obtain event data and metadata.

Your EPL statement can specify a method invocation in the from clause without a join. Such a stand-alone method invocation does not post new events, and may only be queried via the iterator poll API. Your EPL statement may still use variables.

The next statement assumes that a category_var variable has been declared. It polls the GetAssetDescriptions method passing the current value of the category_var EPL variable:

select * from method:MyLookupLib.GetAssetDescriptions(category_var)]

Use the GetEnumerator and GetSafeEnumerator methods on EPStatement to obtain results. The statement does not post events to listeners, it is strictly passive in that sense.

4.16.3. Providing the Method

Your application must provide a type that exposes a public static method. The method must accept the same number and type of parameters as listed in the parameter expression list.

If your method invocation returns either no row or only one row, then the return type of the method can be a native type or a System.Collections.Generic.IDictionary. If your method invocation can return more then one row, then the return type of the method must be an array of native type or an array of Map.

If you are using a native type or an array of native type as the return type, then the class must adhere to property conventions described in this documentation.

If you are using System.Collections.Generic.IDictionary as the return type or an array of Map, then the map should have String-type keys and object values (Map<String, Object>). When using Map as the return type, your application must provide a second method that returns property metadata, as the next section outlines.

Your application method must return either of the following:

  1. A null value or an empty array to indicate an empty result (no rows).

  2. An object or Map to indicate a one-row result, or an array that consists of a single object or Map.

  3. An array of objects or Map instances to return multiple result rows.

As an example, consider the method 'getAssetDescriptions' provided by class 'MyLookupLib' as discussed earlier:

select AssetId, AssetDesc from AssetMoveEvent as asset,
       method:com.mypackage.MyLookupLib.GetAssetDescriptions() as desc 
  where asset.Assetid = desc.Assetid

The 'GetAssetDescriptions' method may return multiple rows and is therefore declared to return an array of the class 'AssetDesc'. The class AssetDesc is a type (not shown here):

public class MyLookupLib {
  ...
  public static AssetDesc[] getAssetDescriptions() {
    ...
    return new AssetDesc[] {...};
  }

The example above specifies the full type name of the type 'MyLookupLib' class in the EPL statement. The package name does not need to be part of the EPL if your application imports the package using the auto-import configuration through the API or XML, as outlined in Section 11.4.5, “Class and package imports”.

4.16.4. Using a Map Return Type

Your application may return System.Collections.Generic.IDictionary or an array of Map from method invocations. If doing so, your application must provide metadata about each row: it must declare the property name and property type of each Map entry of a row. This information allows the engine to perform type checking of expressions used within the statement.

You declare the property names and types of each row by providing a method that returns property metadata. The metadata method must follow these conventions:

  1. The method name providing the property metadata must have same method name appended by the literal Metadata.

  2. The method must have an empty parameter list and must be declared public and static.

  3. The method providing the metadata must return a IDictionary of String property name keys and System.Type property name types (IDictionary<String, Type>).

In the following example, a class 'MyLookupLib' provides a method to return historical data based on asset id and asset code:

select assetId, location, x_coord, y_coord from AssetMoveEvent as asset,
       method:com.mypackage.MyLookupLib.getAssetHistory(assetId, assetCode) as history

A sample implementation of the class 'MyLookupLib' is shown below.

public class MyLookupLib {
  ...
  // For each column in a row, provide the property name and type
  //
  public static Map<String, Class> getAssetHistoryMetadata() {
    Map<String, Class> propertyNames = new HashMap<String, Class>();
    propertyNames.put("location", String.class);
    propertyNames.put("x_coord", Integer.class);
    propertyNames.put("y_coord", Integer.class);
    return propertyNames;
  }
... 
  // Lookup rows based on assetId and assetCode
  // 
  public static Map<String, Object>[] getAssetHistory(String assetId, String assetCode) {
    Map rows = new Map[2];	// this sample returns 2 rows
    for (int i = 0; i < 2; i++) {
      rows[i] = new HashMap();
      rows[i].put("location", "somevalue");
      rows[i].put("x_coord", 100);
      // ... set more values for each row
    }
    return rows;
  }

In the example above, the 'getAssetHistoryMetadata' method provides the property metadata: the names and types of properties in each row. The engine calls this method once per statement to determine event typing information.

The 'getAssetHistory' method returns an array of Map objects that are two rows. The implementation shown above is a simple example. The parameters to the method are the assetId and assetCode properties of the AssetMoveEvent joined to the method. The engine calls this method for each insert and remove stream event in AssetMoveEvent.

To indicate that no rows are found in a join, your application method may return either a null value or an array of size zero.

4.17. Creating and Using Named Windows

A named window is a global data window that can take part in many statement queries, and that can be inserted-into and deleted-from by multiple statements. A named window holds events of the same type or supertype, unless used with a variant stream.

The create window clause declares a new named window. The named window starts up empty unless populated from an existing named window at time of creation. Events must be inserted into the named window using the insert into clause. Events can also be deleted from a named window via the on delete clause.

Events enter the named window by means of insert into clause of a select statement. Events leave a named window either because the expiry policy of the declared data window removes events from the named window, or through statements that use the on delete clause to explicitly delete from a named window.

To query a named window, simply use the window name in the from clause of your statement, including statements that contain subqueries, joins and outer-joins.

A named window may also decorate an event to preserve original events as described in Section 4.10.4, “Decorated Events” and Section 4.17.2.1, “Named Windows Holding Decorated Events”. In addition, columns of a named window are allowed to hold events themselves, as further explained in Section 4.10.5, “Event as a Property” and Section 4.17.2.2, “Named Windows Holding Events As Property”.

To tune subquery performance when the subquery selects from a named window, consider the hints discussed in Section 4.14.7, “Hints Related to Subqueries”.

4.17.1. Creating Named Windows: the Create Window clause

The create window statement creates a named window by specifying a window name and one or more data window views, as well as the type of event to hold in the named window.

There are two syntaxes for creating a named window: The first syntax allows to model a named window after an existing event type or an existing named window. The second syntax is similar to the SQL create-table syntax and provides a list of column names and column types.

A new named window starts up empty. It must be explicitly inserted into by one or more statements, as discussed below. A named window can also be populated at time of creation from an existing named window.

If your application stops or destroys the statement that creates the named window, any consuming statements no longer receive insert or remove stream events. The named window can also not be deleted from after it was stopped or destroyed.

The create window statement posts to listeners any events that are inserted into the named window as new data. The statement posts all deleted events or events that expire out of the data window to listeners as the remove stream (old data). The named window contents can also be iterated on via the pull API to obtain the current contents of a named window.

4.17.1.1. Creation by Modelling after an Existing Type

The benefit of modelling a named window after an existing event type is that event properties can be nested, indexed, mapped or other types that your event objects may provide as properties, including the type of the underlying event itself. Also, using the wildcard (*) operator means your EPL does not need to list each individual property explicitly.

The syntax for creating a named window by modelling the named window after an existing event type, is as follows:

create window window_name.view_specifications 
    [as] [select list_of_properties from] event_type_or_windowname
    [insert [where filter_expression]]

The window_name you assign to the named window can be any identifier. The name should not already be in use as an event type or stream name.

The view_specifications are one or more data window views that define the expiry policy for removing events from the data window. Named windows must explicitly declare a data window view. This is required to ensure that the policy for retaining events in the data window is well defined. To keep all events, use the keep-all view: It indicates that the named window should keep all events and only remove events from the named window that are deleted via the on delete clause. The view specification can only list data window views, derived-value views are not allowed since these don't represent an expiry policy. Data window views are listed in Chapter 9, EPL Reference: Views. View parameterization and staggering are described in Section 4.4.3, “Specifying Views”.

The select clause and list_of_properties are optional. If present, they specify the column names and, implicitly by definition of the event type, the column types of events held by the named window. Expressions other then column names are not allowed in the select list of properties. Wildcards (*) and wildcards with additional properties can also be used.

The event_type_or_windowname is required if using the model-after syntax. It provides the name of the event type of events held in the data window, unless column names and types have been explicitly selected via select. The name of an (existing) other named window is also allowed here. Please find more details in Section 4.17.6, “Populating a Named Window from an Existing Named Window”.

Finally, the insert clause and optional filter_expression are used if the new named windows is modelled after an existing named window, and the data of the existing named window is to be populated, upon time of creation of the new window, from the existing named window. The optional filter_expression can be used to exclude events.

The next statement creates a named window 'AllOrdersNamedWindow' for which the expiry policy is simply to keep all events. Assume that the event type 'OrderMapEventType' has been configured. The named window is to hold events of type 'OrderMapEventType':

create window AllOrdersNamedWindow.win:keepall() as OrderMapEventType

The below sample statement demonstrates the select syntax. It defines a named window in which each row has the three properties 'symbol', 'volume' and 'price'. This named window actively removes events from the window that are older then 30 seconds.

create window OrdersTimeWindow.win:time(30 sec) as 
  select symbol, volume, price from OrderEvent

In an alternate form, the as keyword can be used to rename columns, and constants may occur in the select clause as well:

create window OrdersTimeWindow.win:time(30 sec) as 
  select symbol as sym, volume as vol, price, 1 as alertId from OrderEvent

4.17.1.2. Creation By Defining Columns Names and Types

The second syntax for creating a named window is by supplying column names and types:

create window window_name.view_specifications [as] (column_name column_type 
  [,column_name column_type [,...])

The column_name is an identifier providing the event property name. The column_type is also required for each column. Valid column types are listed in Section 4.20.1, “Creating Variables: the Create Variable clause” and are the same as for variable types.

For attributes that are array-type append [] (left and right brackets).

The next statement creates a named window:

create window SecurityEvent.win:time(30 sec) 
    (ipAddress string, userId String, numAttempts int, properties String[])

Named window columns can hold events by declaring the column type as the event type name. Array-type is combination with event-type is also supported.

The next two statements declare an event type and create a named window with a column of the defined event type:

create schema SecurityData (name String, roles String[])
create window SecurityEvent.win:time(30 sec) 
    (ipAddress string, userId String, secData SecurityData, historySecData SecurityData[])

4.17.1.3. Dropping or Removing Named Windows

There is no syntax to drop or remove a named window.

The Dispose method on the EPStatement that created the named window removes the named window. However the implicit event type associated with the named window remains active since further statements may continue to use that type. Therefore a named window of the same name can only be created again if the type information matches the prior declaration for a named window.

4.17.2. Inserting Into Named Windows

The insert into clause inserts events into named windows. Your application must ensure that the column names and types match the declared column names and types of the named window to be inserted into.

For inserting into a named window and for simulateously checking if the inserted row already exists in the named window or for atomic update-insert operation on a named window, consider using on-merge as described in Section 4.17.9, “Triggered Upsert using the On-Merge Clause”. On-merge is similar to the SQL merge clause and provides what is known as an "Upsert" operation: Update existing events or if no existing event(s) are found then insert a new event, all in one atomic operation provided by a single EPL statement.

In this example we first create a named window using some of the columns of an OrderEvent event type:

create window OrdersWindow.win:keepall() as select symbol, volume, price from OrderEvent

The insert into the named window selects individual columns to be inserted:

insert into OrdersWindow(symbol, volume, price) select name, count, price from FXOrderEvent

An alternative form is shown next:

insert into OrdersWindow select name as symbol, vol as volume, price from FXOrderEvent

Following above statement, the engine enters every FXOrderEvent arriving into the engine into the named window 'OrdersWindow'.

The following EPL statements create a named window for an event type backed by a type and insert into the window any 'OrderEvent' where the symbol value is IBM:

create window OrdersWindow.win:time(30) as com.mycompany.OrderEvent
insert into OrdersWindow select * from com.mycompany.OrderEvent(symbol='IBM')

The last example adds one column named 'derivedPrice' to the 'OrderEvent' type by specifying a wildcard, and uses a user-defined function to populate the column:

create window OrdersWindow.win:time(30) as select *, price as derivedPrice from OrderEvent
insert into OrdersWindow select *, MyFunc.func(price, percent) as derivedPrice from OrderEvent

Event representations based on base classes or interfaces, and subclasses or implementing classes, are compatible as these statements show:

// create a named window for the base class
create window OrdersWindow.std:unique(name) as select * from ProductBaseEvent
// The ServiceProductEvent class subclasses the ProductBaseEvent class
insert into OrdersWindow select * from ServiceProductEvent
// The MerchandiseProductEvent class subclasses the ProductBaseEvent class
insert into OrdersWindow select * from MerchandiseProductEvent

To avoid duplicate events inserted in a named window and atomically check if a row already exists, use on-merge as outlined in Section 4.17.9, “Triggered Upsert using the On-Merge Clause”. An example:

on ServiceProductEvent as spe merge OrdersWindow as win
where win.id = spe.id when not matched then insert select *

4.17.2.1. Named Windows Holding Decorated Events

Decorated events hold an underlying event and add additional properties to the underlying event, as described further in Section 4.10.4, “Decorated Events”.

Here we create a named window that decorates OrderEvent events by adding an additional property named priceTotal to each OrderEvent. A matching insert into statement is also part of the sample:

create window OrdersWindow.win:time(30) as select *, price as priceTotal from OrderEvent
insert into OrdersWindow select *, price * unit as priceTotal from ServiceOrderEvent

The property type of the additional priceTotal column is the property type of the existing price property of OrderEvent.

4.17.2.2. Named Windows Holding Events As Property

Columns in a named window may also hold an event itself. More information on the insert into clause providing event columns is in Section 4.10.5, “Event as a Property”.

The next sample creates a named window that specifies two columns: A column that holds an OrderEvent, and a column by name priceTotal. A matching insert into statement is also part of the sample:

create window OrdersWindow.win:time(30) as select this, price as priceTotal from OrderEvent
insert into OrdersWindow select order, price * unit as priceTotal  
from ServiceOrderEvent as order

Note that the this proprerty must exist on the event and must return the event type itself (native events only). The property type of the additional priceTotal column is the property type of the existing price property.

4.17.3. Selecting From Named Windows

A named window can be referred to by any statement in the from clause of the statement. Filter criteria can also be specified. Additional views may be used onto named windows however such views cannot include data window views.

A statement selecting all events from a named window 'AllOrdersNamedWindow' is shown next. The named window must first be created via the create window clause before use.

select * from AllOrdersNamedWindow

The statement as above simply receives the unfiltered insert stream of the named window and reports that stream to its listeners. The GetEnumerator method returns returns all events in the named window, if any.

If your application desires to obtain the events removed from the named window, use the rstream keyword as this statement shows:

select rstream * from AllOrdersNamedWindow

The next statement derives an average price per symbol from all events posted by a named window:

select symbol, avg(price) from AllOrdersNamedWindow group by symbol

Your application may create a consuming statement such as above on an empty named window, or your application may create the above statement on an already filled named window. The engine provides correct results in either case: At the time of statement creation the Esper engine internally initializes the consuming statement from the current named window, also taking your declared filters into consideration. Thus, your statement deriving data from a named window does not start empty if the named window already holds one or more events. A consuming statement also sees the remove stream of an already populated named window, if any.

If you require a subset of the data in the named window, you can specify one or more filter expressions onto the named window as shown here:

select symbol, avg(price) from AllOrdersNamedWindow(sector='energy') group by symbol

By adding a filter to the named window, the aggregation and grouping as well as any views that may be declared onto to the named window receive a filtered insert and remove stream. The above statement thus outputs, continuously, the average price per symbol for all orders in the named window that belong to a certain sector.

A side note on variables in filters filtering events from named windows: The engine initializes consuming statements at statement creation time and changes aggregation state continuously as events arrive. If the filter criteria contain variables and variable values changes, then the engine does not re-evaluate or re-build aggregation state. In such a case you may want to place variables in the having clause which evaluates on already-built aggregation state.

The following example further declares a view into the named window. Such a view can be a plug-in view or one of the built-in views, but cannot be a data window view (with the exception of the std:groupwin grouped-window view which is allowed).

select * from AllOrdersNamedWindow(volume>0, price>0).mycompany:mypluginview()

Data window views cannot be used onto named windows since named windows post insert and remove streams for the events entering and leaving the named window, thus the expiry policy and batch behavior are well defined by the data window declared for the named window. For example, the following is not allowed and fails at time of statement creation:

// not a valid statement
select * from AllOrdersNamedWindow.win:time(30 sec)

4.17.4. Triggered Select on Named Windows: the On Select clause

The on select clause performs a one-time, non-continuous query on a named window every time a triggering event arrives or a triggering pattern matches. The query can consider all events in the named window, or only events that match certain criteria, or events that correlate with an arriving event or a pattern of arriving events.

The syntax for the on select clause is as follows:

on event_type[(filter_criteria)] [as stream_name]
[insert into insert_into_def]
select select_list
from window_name [as stream_name]
[where criteria_expression]
[group by grouping_expression_list]
[having grouping_search_conditions]
[order by order_by_expression_list]

The event_type is the name of the type of events that trigger the query against the named window. It is optionally followed by filter_criteria which are filter expressions to apply to arriving events. The optional as keyword can be used to assign an stream name. Patterns or named windows can also be specified in the on clause, see the samples in Section 4.17.8.1, “Using Patterns in the On Delete Clause”.

The insert into clause works as described in Section 4.10, “Merging Streams and Continuous Insertion: the Insert Into Clause”. The select clause is described in Section 4.3, “Choosing Event Properties And Events: the Select Clause”. For all clauses the semantics are equivalent to a join operation: The properties of the triggering event or events are available in the select clause and all other clauses.

The window_name in the from clause is the name of the named window to select events from. The as keyword is also available to assign a stream name to the named window. The as keyword is helpful in conjunction with wildcard in the select clause to select named window events via the syntax select streamname.* .

The optional where clause contains a criteria_expression that correlates the arriving (triggering) event to the events to be considered from the named window. The criteria_expression may also simply filter for events in the named window to be considered by the query.

The group by clause, the having clause and the order by clause are all optional and work as described in earlier chapters.

The similarities and differences between an on select clause and a regular or outer join are as follows:

  1. A join is evaluated when any of the streams participating in the join have new events (insert stream) or events leaving data windows (remove stream). A join is therefore bi-directional or multi-directional. However, the on select statement has one triggering event or pattern that causes the query to be evaluated and is thus uni-directional.

  2. The query within the on select statement is not continuous: It executes only when a triggering event or pattern occurs. Aggregation and groups are computed anew considering the contents of the named window at the time the triggering event arrives.

The iterator of the EPStatement object representing the on select clause returns the last batch of selected events in response to the last triggering event, or null if the last triggering event did not select any rows.

For correlated queries that correlate triggering events with events held by a named window, Esper internally creates efficient indexes to enable high performance querying of events. It analyzes the where clause to build one or more indexes for fast lookup in the named window based on the properties of the triggering event.

The next statement demonstrates the concept. Upon arrival of a QueryEvent event the statement selects all events in the 'OrdersNamedWindow' named window:

on QueryEvent select win.* from OrdersNamedWindow as win

The engine executes the query on arrival of a triggering event, in this case a QueryEvent. It posts the query results to any listeners to the statement, in a single invocation, as the new data array. By prefixing the wildcard (*) selector with the stream name, the select clause returns only events of the named window and does not also return triggering events.

The where clause filters and correlates events in the named window with the triggering event, as shown next:

on QueryEvent(volume>0) as query
select query.symbol, query.volume, win.symbol  from OrdersNamedWindow as win
where win.symbol = query.symbol

Upon arrival of a QueryEvent, if that event has a value for the volume property that is greater then zero, the engine executes the query. The query considers all events currently held by the 'OrdersNamedWindow' that match the symbol property value of the triggering QueryEvent event. The engine then posts query results to the statement's listeners.

Aggregation, grouping and ordering of results are possible as this example shows:

on QueryEvent as queryEvent
select symbol, sum(volume) from OrdersNamedWindow as win
group by symbol
having volume > 0
order by symbol

The above statement outputs the total volume per symbol for those groups where the sum of the volume is greater then zero, ordered by symbol ascending. The engine computes and posts the output based on the current contents of the 'OrdersNamedWindow' named window considering all events in the named window, since the query does not have a where clause.

When using wildcard (*) to select from streams in an on-select clause, each stream, that is the the triggering stream and the selected-upon named window, are selected, similar to a join. Therefore your wildcard select returns two columns: the triggering event and the selection result event, for each row.

on QueryEvent as queryEvent
select * from OrdersNamedWindow as win

The query above returns a queryEvent column and a win column for each event. If only a single stream's event is desired in the result, use select win.* instead.

To trigger an on-select when an update to the selected named window occurs or when the triggering event is the same event that is being inserted into the named window, specify the named window name as the event type.

The next query fires the select for every change to the named window OrdersNamedWindow:

on OrdersNamedWindow as trig 
select onw.symbol, sum(onw.volume) 
from OrdersNamedWindow as onw 
where onw.symbol = trig.symbol

4.17.5. Triggered Playback from Named Windows: the On Insert clause

The on insert clause is an on select clause as described in the prior chapter with the addition of an insert into clause.

Similar to the on select clause, the engine executes the query when a triggering event arrives. It then provides the query results as an event stream to further statements. It populates the event stream that is named in the insert into clause.

The statement below provides the query results to any consumers of the MyOrderStream, upon arrival of a QueryEvent event:

on QueryEvent as query
insert into MyOrderStream
select win.* from OrdersNamedWindow as win

Here is a sample consuming statement of the MyOrderStream. The statement further filters the events provided by the on insert statement by user id and reports a total of volume per symbol:

select symbol, sum(volume) from MyOrderStream(userId='user1') group by symbol

4.17.6. Populating a Named Window from an Existing Named Window

Your EPL statement may specify the name of an existing named window when creating a new named window, and may use the insert keyword to indicate that the new named window is to be populated from the events currently held by the existing named window.

For example, and assuming the named window OrdersNamedWindow already exists, this statement creates a new named window ScratchOrders and populates all orders in OrdersNamedWindow into the new named window:

create window ScratchOrders.win:keepall() as OrdersNamedWindow insert

The where keyword is also available to perform filtering, for example:

create window ScratchBuyOrders.win:time(10) as OrdersNamedWindow insert where side = 'buy'

4.17.7. Updating Named Windows: the On Update clause

An on update clause updates events held by a named window. The clause can be used to update all events, or only events that match certain criteria, or events that correlate with an arriving event or a pattern of arriving events.

For updating a named window and for simulateously checking if the updated row exists in the named window or for atomic update-insert operation on a named window, consider using on-merge as described in Section 4.17.9, “Triggered Upsert using the On-Merge Clause”. On-merge is similar to the SQL merge clause and provides what is known as an "Upsert" operation: Update existing events or if no existing event(s) are found then insert a new event, all in one atomic operation provided by a single EPL statement.

The syntax for the on update clause is as follows:

on event_type[(filter_criteria)] [as stream_name]
update window_name [as stream_name]
set property_name = expression [, property_name = expression [,...]]
[where criteria_expression]

The event_type is the name of the type of events that trigger an update of rows in a named window. It is optionally followed by filter_criteria which are filter expressions to apply to arriving events. The optional as keyword can be used to assign a name for use in expressions and the where clause. Patterns and named windows can also be specified in the on clause.

The window_name is the name of the named window to update events. The as keyword is also available to assign a name to the named window.

The comma-separated list of property names and expressions set the value of one or more properties. Subqueries may by part of expressions however aggregation functions and the prev or prior function may not be used in expressions.

The optional where clause contains a criteria_expression that correlates the arriving (triggering) event to the events to be updated in the named window. The criteria_expression may also simply filter for events in the named window to be updated.

The iterator of the EPStatement object representing the on update clause can also be helpful: It returns the last batch of updated events in response to the last triggering event, in any order, or null if the last triggering event did not update any rows.

Statements that reference the named window receive the new event in the insert stream and the event prior to the update in the remove stream.

Let's look at a couple of examples. In the simplest form, this statement updates all events in the named window 'AllOrdersNamedWindow' when any 'UpdateOrderEvent' event arrives, setting the price property to zero for all events currently held by the named window:

on UpdateOrderEvent update AllOrdersNamedWindow set price = 0

This example adds a where clause to the example above. Upon arrival of a triggering 'ZeroVolumeEvent', the statement updates prices on any orders that have a volume of zero or less:

on ZeroVolumeEvent update AllOrdersNamedWindow set price = 0 where volume <= 0

The next example shows a more complete use of the syntax, and correlates the triggering event with events held by the named window:

on NewOrderEvent(volume>0) as myNewOrders
update AllOrdersNamedWindow as myNamedWindow 
set price = myNewOrders.price
where myNamedWindow.symbol = myNewOrders.symbol

In the above sample statement, only if a 'NewOrderEvent' event with a volume greater then zero arrives does the statement trigger. Upon triggering, all events in the named window that have the same value for the symbol property as the triggering 'NewOrderEvent' event are then updated (their price property is set to that of the arriving event). The statement also showcases the as keyword to assign a name for use in the where expression.

For correlated queries (as above) that correlate triggering events with events held by a named window, Esper internally creates efficient indexes to enable high performance update of events.

Your application can subscribe a listener to your on update statements to determine update events. The statement post any events that are updated to all listeners attached to the statement as new data, and the events prior to the update as old data. Upon iteration, the statement provides the last update event, if any.

The following example shows the use of tags and a pattern. It sets the price value of orders to that of either a 'FlushOrderEvent' or 'OrderUpdateEvent' depending on which arrived:

on pattern [every ord=OrderUpdateEvent(volume>0) or every flush=FlushOrderEvent] 
update AllOrdersNamedWindow as win
set price = case when ord.price is null then flush.price else ord.price end
where ord.id = win.id or flush.id = win.id

The following restrictions apply:

  1. Each property to be updated must be writable.

  2. For underlying event representations that are CLR objects, a event object class must be Serializable as discussed in Section 4.22.1, “Immutability and Updates” and must provide setter methods for updated properties.

  3. When using an XML underlying event type, event properties in the XML document representation are not available for update.

  4. Nested, indexed and mapped properties are not supported for update. Revision event types and variant streams may also not be updated.

4.17.8. Deleting From Named Windows: the On Delete clause

An on delete clause removes events from a named window. The clause can be used to remove all events, or only events that match certain criteria, or events that correlate with an arriving event or a pattern of arriving events.

The syntax for the on delete clause is as follows:

on event_type[(filter_criteria)] [as stream_name]
delete from window_name [as stream_name]
[where criteria_expression]

The event_type is the name of the type of events that trigger removal from the named window. It is optionally followed by filter_criteria which are filter expressions to apply to arriving events. The optional as keyword can be used to assign a name for use in the where clause. Patterns and named windows can also be specified in the on clause as described in the next section.

The window_name is the name of the named window to delete events from. The as keyword is also available to assign a name to the named window.

The optional where clause contains a criteria_expression that correlates the arriving (triggering) event to the events to be removed from the named window. The criteria_expression may also simply filter for events in the named window to be removed from the named window.

The iterator of the EPStatement object representing the on delete clause can also be helpful: It returns the last batch of deleted events in response to the last triggering event, in any order, or null if the last triggering event did not remove any rows.

Let's look at a couple of examples. In the simplest form, this statement deletes all events from the named window 'AllOrdersNamedWindow' when any 'FlushOrderEvent' arrives:

on FlushOrderEvent delete from AllOrdersNamedWindow

This example adds a where clause to the example above. Upon arrival of a triggering 'ZeroVolumeEvent', the statement removes from the named window any orders that have a volume of zero or less:

on ZeroVolumeEvent delete from AllOrdersNamedWindow where volume <= 0

The next example shows a more complete use of the syntax, and correlates the triggering event with events held by the named window:

on NewOrderEvent(volume>0) as myNewOrders
delete from AllOrdersNamedWindow as myNamedWindow 
where myNamedWindow.symbol = myNewOrders.symbol

In the above sample statement, only if a 'NewOrderEvent' event with a volume greater then zero arrives does the statement trigger. Upon triggering, all events in the named window that have the same value for the symbol property as the triggering 'NewOrderEvent' event are then removed from the named window. The statement also showcases the as keyword to assign a name for use in the where expression.

For correlated queries (as above) that correlate triggering events with events held by a named window, Esper internally creates efficient indexes to enable high performance removal of events especially from named windows that hold large numbers of events.

Your application can subscribe a listener to your on delete statements to determine removed events. The statement post any events that are deleted from a named window to all listeners attached to the statement as new data. Upon iteration, the statement provides the last deleted event, if any.

4.17.8.1. Using Patterns in the On Delete Clause

By means of patterns the on delete clause and on select clause (described below) can look for more complex conditions to occur, possibly involving multiple events or the passing of time. The syntax for on delete with a pattern expression is show next:

on pattern [pattern_expression] [as stream_name]
delete from window_name [as stream_name]
[where criteria_expression]

The pattern_expression is any pattern that matches zero or more arriving events. Tags can be used to name events in the pattern and can occur in the optional where clause to correlate to events to be removed from a named window.

In the next example the triggering pattern fires every 10 seconds. The effect is that every 10 seconds the statement removes from 'MyNamedWindow' all rows:

on pattern [every timer:interval(10 sec)] delete from MyNamedWindow

The following example shows the use of tags in a pattern:

on pattern [every ord=OrderEvent(volume>0) or every flush=FlushOrderEvent] 
delete from OrderWindow as win
where ord.id = win.id or flush.id = win.id

The pattern above looks for OrderEvent events with a volume value greater then zero and tags such events as 'ord'. The pattern also looks for FlushOrderEvent events and tags such events as 'flush'. The where clause deletes from the 'OrderWindow' named window any events that match in the value of the 'id' property either of the arriving events.

4.17.9. Triggered Upsert using the On-Merge Clause

The on merge clause is similar to the SQL merge clause. It provides what is known as an "Upsert" operation: Update existing events or if no existing event(s) are found then insert a new event, all in an atomic operation provided by a single EPL statement.

The syntax for the on merge clause is as follows:

on event_type[(filter_criteria)] [as stream_name]
merge [into] window_name [as stream_name]
[where criteria_expression]
  when [not] matched [ and search_condition ]
    then [
      insert [ (property_name [, property_name] [,...]) ] 
          select select_expression [, select_expression[,...]]
      |
      update set property_name = expression [, property_name = expression [,...]]
      |
      delete
    ]  
  [when ...  then ... [...]] 

The event_type is the name of the type of events that trigger the merge. It is optionally followed by filter_criteria which are filter expressions to apply to arriving events. The optional as keyword can be used to assign a name for use in the where clause. Patterns and named windows can also be specified in the on clause as described in prior sections.

The window_name is the name of the named window to insert, update or delete events. The as keyword is also available to assign a name to the named window.

The optional where clause contains a criteria_expression that correlates the arriving (triggering) event to the events to be considered of the named window.

Following the where clause are one or more when matched or when not matched clauses in any order. Each may have an additional search condition associated. It follows the then keyword followed by either an insert, update or delete keyword.

After when not matched only the insert keyword and action are available. After insert follows an optional list of columns inserted, the select keyword and one or more select clause expressions. The wildcard (*) is available in the select clause as well.

After when matched only the update or delete keywords and actions are available. After update follows the set keyword and one or more assignment pairs.

When according to the where-clause criteria_expression the engine finds no events in the named window that match the condition, the engine evaluates each when not matched clause. If the optional search condition returns true or no search condition was provided then the engine performs the insert action as specified.

When according to the where-clause criteria_expression the engine finds one or more events in the named window that match the condition, the engine evaluates each when matched clause. If the optional search condition returns true or no search condition was provided the engine performs an update or delete action as specified.

The engine executes when matched and when not matched in the order specified. If the optional search condition returns true or no search condition was specified then the engine takes the associated action. When the action completed the engine proceeds to the next matching event, if any. After completing all matching events the engine continues to the next triggering event if any.

In the first example we declare a schema that provides a product id and that holds a total price:

create schema ProductTotalRec as (productId string, totalPrice double)

We create a named window that holds a row for each unique product:

create window ProductWindow.std:unique(productId) as ProductTotalRec

The events for this example are order events that hold an order id, product id, price, quantity and deleted-flag declared by the next schema:

create schema OrderEvent as (orderId string, productId string, price double, 
    quantity int, deletedFlag boolean)

The following EPL statement utilizes on-merge to total up the price for each product based on arriving order events:

on OrderEvent oe
  merge ProductWindow pw
  where pw.productId = oe.productId
  when matched
    then update set totalPrice = totalPrice + oe.price
  when not matched 
    then insert select productId, price as totalPrice

In the above example, when an order event arrives, the engine looks up in the product named window the matching row or rows for the same product id as the arriving event. In this example the engine always finds no row or one row as the product named window is declared with a unique data window based on product id. If the engine finds a row in the named window, it performs the update action adding up the price as defined under when matched. If the engine does not find a row in the named window it performs the insert action as defined under when not matched, inserting a new row.

The insert keyword may be followed by a list of columns as shown in this EPL snippet:

// equivalent to the insert shown in the last 2 lines in above EPL
...when not matched 
    then insert(productId, totalPrice) select productId, price

The second example demonstrates the use of a select-clause with wildcard, a search condition and the delete keyword. It creates a named window that holds order events and employs on-merge to insert order events for which no corresponding order id was found, update quantity to the quantity provided by the last arriving event and delete order events that are marked as deleted:

create window OrderWindow.win:keepall() as OrderEvent
on OrderEvent oe
  merge OrderWindow pw
  where pw.orderId = oe.orderId
  when not matched 
    then insert select *
  when matched and oe.deletedFlag=true
    then delete
  when matched
    then update set pw.quantity = oe.quantity, pw.price = oe.price

In the above example the oe.deletedFlag=true search condition instructs the engine to take the delete action only if the deleted-flag is set.

For correlated queries (as above) that correlate triggering events with events held by a named window, Esper internally creates efficient indexes to enable high performance update and removal of events especially from named windows that hold large numbers of events.

Your application can subscribe a listener to on merge statements to determine inserted, updated and removed events. Statements post any events that are inserted to, updated or deleted from a named window to all listeners attached to the statement as new data and removed data. Upon iteration, the statement provides the last inserted events, if any.

4.17.10. Explicitly Indexing Named Windows

You may explicitly create an index on a named window. When executing on-demand (fire-and-forget) queries as described in Section 10.4.3, “On-Demand Snapshot Query Execution” using prepared or unprepared queries, an explicit index can help speed query performance.

You do not need to explicitly create an index unless using on-demand (fire-and-forget) queries. Therefore you do not need to explicitly create an index for on-select, on-update, on-delete as well as subqueries or consuming statements against named windows.

Please use the following syntax to create an explicit index on a named window:

create index index_name on named_window_name (property[, property] [,...] )

The index_name is the name assigned to the index. The name uniquely identifies the index and is used in engine logging.

The named_window_name is the name of an existing named window. If the named window has data already, the engine builds an index for the data in the named window.

The list of property names are the properties of events within the named window to include in the index.

We list a few example EPL statements next that create a named window and create a single index:

// create a named window
create window UserProfileWindow.win:time(1 hour) select * from UserProfile
// create an index for the user id property only
create index UserProfileIndex on UserProfileWindow(userId)

Next, execute a on-demand fire-and-forget query as shown below, herein we use the prepared version to demonstrate:

String query = "select * from UserProfileWindow where userId='Joe'";
EPOnDemandPreparedQuery prepared = epRuntime.prepareQuery(query);
// query performance excellent in the face of large number of rows
EPOnDemandQueryResult result = prepared.execute();
// ...later ...
prepared.execute();	// execute a second time

The engine builds a hash code -based index useful for direct comparison via equals (=), filter expressions that look for ranges or use in, between do not benefit from the hash-based index. Named windows that hold a revision event type or a variant stream event type may not be indexed.

4.17.11. Versioning and Revision Event Type Use with Named Windows

As outlined in Section 2.9, “Updating, Merging and Versioning Events”, revision event types process updates or new versions of events held by a named window.

A revision event type is simply one or more existing pre-configured event types whose events are related, as configured by static configuration, by event properties that provide same key values. The purpose of key values is to indicate that arriving events are related: An event amends, updates or adds properties to an earlier event that shares the same key values. No additional EPL is needed when using revision event types for merging event data.

Revision event types can be useful in these situations:

  1. Some of your events carry only partial information that is related to a prior event and must be merged together.

  2. Events arrive that add additional properties or change existing properties of prior events.

  3. Events may carry properties that have null values or properties that do no exist (for example events backed by Map or XML), and for such properties the earlier value must be used instead.

To better illustrate, consider a revision event type that represents events for creation and updates to user profiles. Lets assume the user profile creation events carry the user id and a full profile. The profile update events indicate only the user id and the individual properties that actually changed. The user id property shall serve as a key value relating profile creation events and update events.

A revision event type must be configured to instruct the engine which event types participate and what their key properties are. Configuration is described in Section 11.4.22, “Revision Event Type” and is not shown here.

Assume that an event type UserProfileRevisions has been configured to hold profile events, i.e. creation and update events related by user id. This statement creates a named window to hold the last 1 hour of current profiles per user id:

create window UserProfileWindow.win:time(1 hour) select * from UserProfileRevisions
insert into UserProfileWindow select * from UserProfileCreation
insert into UserProfileWindow select * from UserProfileUpdate

In revision event types, the term base event is used to describe events that are subject to update. Events that update, amend or add additional properties to base events are termed delta events. In the example, base events are profile creation events and delta events are profile update events.

Base events are expected to arrive before delta events. In the case where a delta event arrives and is not related by key value to a base event or a revision of the base event currently held by the named window the engine ignores the delta event. Thus, considering the example, profile update events for a user id that does not have an existing profile in the named window are not applied.

When a base or delta event arrives, the insert and remove stream output by the named window are the current and the prior version of the event. Let's come back to the example. As creation events arrive that are followed by update events or more creation events for the same user id, the engine posts the current version of the profile as insert stream (new data) and the prior version of the profile as remove stream (old data).

Base events are also implicitly delta events. That is, if multiple base events of the same key property values arrive, then each base event provides a new version. In the example, if multiple profile creation events arrive for the same user id then new versions of the current profile for that user id are output by the engine for each base event, as it does for delta events.

The expiry policy as specified by view definitions applies to each distinct key value, or multiple distinct key values for composite keys. An expiry policy re-evaluates when new versions arrive. In the example, user profile events expire from the time window when no creation or update event for a given user id has been received for 1 hour.

Several strategies are available for merging or overlaying events as the configuration chapter describes in greater detail.

Any of the Map, XML and native event representations as well as plug-in event representations may participate in a revision event type. For example, profile creation events could be native events, while profile update events could be System.Collections.Generic.IDictionary events.

Delta events may also add properties to the revision event type. For example, one could add a new event type with security information to the revision event type and such security-related properties become available on the resulting revision event type.

The following restrictions apply to revision event types:

  • Nested properties are only supported for the native event representation. Nested properties are not individually versioned; they are instead versioned by the containing property.

  • Dynamic, indexed and mapped properties are only supported for nested properties and not as properties of the revision event type itself.

4.18. Declaring an Event Type: Create Schema

EPL allows declaring an event type via the create schema clause and also by means of the static or runtime configuration API AddEventType functions. The term schema and event type has the same meaning in EPL.

Your application can declare an event type by providing the property names and types or by providing a class name. Your application may also declare a variant stream schema.

When using the create schema syntax to declare an event type, the engine automatically removes the event type when there are no started statements referencing the event type, including the statement that declared the event type. When using the configuration API, the event type stays cached even if there are no statements that refer to the event type and until explicitly removed via the runtime configuration API.

4.18.1. Declare an Event Type by Providing Names and Types

The synopsis of the create schema syntax providing property names and types is:

create schema schema_name [as] (property_name property_type [,property_name property_type [,...])
  [inherits inherited_event_type[, inherited_event_type] [,...]]

The property_name is an identifier providing the event property name. The property_type is also required for each property. Valid property types are listed in Section 4.20.1, “Creating Variables: the Create Variable clause” and in addition include:

  1. Any type name, fully-qualified or the simple class name if imports are configured.

  2. Add left and right square brackets [] to any type to denote an array-type event property.

  3. Use an event type name as a property type.

The optional inherits keywords may be used to provide a list of event types that are supertypes to the declared type.

A few example event type declarations follow:

// Declare type SecurityEvent
create schema SecurityEvent as (ipAddress string, userId String, numAttempts int)
			
// Declare type AuthorizationEvent with the roles property being an array of String 
// and the hostinfo property being an object
create schema AuthorizationEvent(group String, roles String[], hostinfo com.mycompany.HostNameInfo)

// Declare type CompositeEvent in which the innerEvents property is an array of SecurityEvent
create schema CompositeEvent(group String, innerEvents SecurityEvent[])

// Declare type WebPageVisitEvent that inherits all properties from PageHitEvent
create schema WebPageVisitEvent(userId String) inherits PageHitEvent

4.18.2. Declare an Event Type by Providing a Class Name

When using native types as the underlying event representation your application may simply provide the class name:

create schema schema_name [as] class_name

The class_name must be a fully-qualified class name (including the package name) if imports are not configured. If you application configures imports then the simple class name suffices without package name.

The next example statements declare an event type based on a class:

// Shows the use of a fully-qualified class name to declare the LoginEvent event type
create schema LoginEvent as com.mycompany.LoginValue

// When the configuration includes imports, the declaration does not need a package name
create schema LogoutEvent as SignoffValue

4.18.3. Declare a Variant Stream

A variant stream is a predefined stream into which events of multiple disparate event types can be inserted. Please see Section 4.10.3, “Merging Disparate Types of Events: Variant Streams” for rules regarding property visibility and additional information.

The synopsis is:

create variant schema schema_name [as] eventtype_name|* [, eventtype_name|*] [,...]

Provide the variant keyword to declare a variant stream.

The '*' wildcard character declares a variant stream that accepts any type of event inserted into the variant stream.

Provide eventtype_name if the variant stream should hold events of the given type only. When using insert into to insert into the variant stream the engine checks to ensure the inserted event type or its supertypes match the required event type.

A few examples are shown below:

// Create a variant stream that accepts only LoginEvent and LogoutEvent event types
create variant schema SecurityVariant as LoginEvent, LogoutEvent

// Create a variant stream that accepts any event type
create variant schema AnyEvent as *

4.19. Splitting and Duplicating Streams

EPL offers a convenient syntax to splitting and duplicating events into multiple streams, and for receiving unmatched events among a set of filter criteria.

You may define a triggering event or pattern in the on-part of the statement followed by multiple insert into, select and where clauses.

The synopsis is:

on event_type[(filter_criteria)] [as stream_name]
insert into insert_into_def select select_list [where condition]
[insert into insert_into_def select select_list [where condition]]
[insert into...]
[output first | all]

The event_type is the name of the type of events that trigger the split stream. It is optionally followed by filter_criteria which are filter expressions to apply to arriving events. The optional as keyword can be used to assign a stream name. Patterns and named windows can also be specified in the on clause.

Following the on-clause is one or more insert into clauses as described in Section 4.10, “Merging Streams and Continuous Insertion: the Insert Into Clause” and select clauses as described in Section 4.3, “Choosing Event Properties And Events: the Select Clause”.

Each select clause may be followed by a where clause containing a condition. If the condition is true for the event, the engine transforms the event according to the select clause and inserts it into the corresponding stream.

At the end of the statement can be an optional output clause. By default the engine inserts into the first stream for which the where clause condition matches if one was specified, starting from the top. If you specify the output all keywords, then the engine inserts into each stream (not only the first stream) for which the where clause condition matches or that do not have a where clause.

If, for a given event, none of the where clause conditions match, the statement listener receives the unmatched event. The statement listener only receives unmatched events and does not receive any transformed or inserted events. The GetEnumerator method to the statement returns no events.

In the below sample statement, the engine inserts each OrderEvent into the LargeOrders stream if the order quantity is 100 or larger, or into the SmallOrders stream if the order quantity is smaller then 100:

on OrderEvent 
  insert into LargeOrders select * where orderQty >= 100
  insert into SmallOrders select *

The next example statement adds a new stream for medium-sized orders. The new stream receives orders that have an order quantity between 20 and 100:

on OrderEvent 
  insert into LargeOrders select orderId, customer where orderQty >= 100
  insert into MediumOrders select orderId, customer where orderQty between 20 and 100
  insert into SmallOrders select orderId, customer where orderQty > 0

As you may have noticed in the above statement, orders that have an order quantity of zero don't match any of the conditions. The engine does not insert such order events into any stream and the listener to the statement receives these unmatched events.

By default the engine inserts into the first insert into stream without a where clause or for which the where clause condition matches. To change the default behavior and insert into all matching streams instead (including those without a where clause), the output all keywords may be added to the statement.

The sample statement below shows the use of the output all keywords. The statement populates both the LargeOrders stream with large orders as well as the VIPCustomerOrders stream with orders for certain customers based on customer id:

on OrderEvent 
  insert into LargeOrders select * where orderQty >= 100
  insert into VIPCustomerOrders select * where customerId in (1001, 1002)
  output all

Since the output all keywords are present, the above statement inserts each order event into either both streams or only one stream or none of the streams, depending on order quantity and customer id of the order event. The statement delivers order events not inserted into any of the streams to the listeners and/or subscriber to the statement.

The following limitations apply to split-stream statements:

  1. Aggregation functions and the prev and prior operators are not available in conditions and the select-clause.

4.20. Variables

A variable is a scalar, object or event value that is available for use in all statements including patterns. Variables can be used in an expression anywhere in a statement as well as in the output clause for output rate limiting.

Variables must first be declared or configured before use, by defining each variable's type and name. Variables can be created via the create variable syntax or declared by runtime or static configuration. Variables can be assigned new values by using the on set syntax or via the SetVariableValue methods on EPRuntime. The EPRuntime also provides method to read variable values.

When declaring a class-type or an event type variable you may read or set individual properties within the same variable.

The engine guarantees consistency and atomicity of variable reads and writes on a statement-level (this is a soft guarantee, see below). Variables are optimized for fast read access and are also multithread-safe.

Variables can also be removed, at runtime, by destroying all referencing statements including the statement that created the variable, or by means of the runtime configuration API.

4.20.1. Creating Variables: the Create Variable clause

The create variable syntax creates a new variable by defining the variable type and name. In alternative to the syntax, variables can also be declared in the runtime and engine configuration options.

The synopsis for creating a variable is as follows:

create variable variable_type variable_name [ = assignment_expression ]

The variable_type can be any of the following:

variable_type
	:  string
	|  char 
	|  character
	|  bool 
	|  boolean
	|  byte
	|  short 
	|  int 
	|  integer 
	|  long 
	|  double
	|  float
	|  object
	|  class_name
	|  event_type_name

All variable types accept null values. The object type is for an untyped variable that can be assigned any value. You can provide a fully-qualified class name to declare a variable of that class type. You can also supply the name of an event type to declare a variable that holds an event of that type.

The variable_name is an identifier that names the variable. The variable name should not already be in use by another variable.

The assignment_expression is optional. Without an assignment expression the initial value for the variable is null. If present, it supplies the initial value for the variable.

The EPStatement object of the create variable statement provides access to variable values. The pull API methods GetEnumerator and GetSafeEnumerator return the current variable value. Listeners to the create variable statement subscribe to changes in variable value: the engine posts new and old value of the variable to all listeners when the variable value is updated by an on set statement.

The example below creates a variable that provides a threshold value. The name of the variable is var_threshold and its type is long. The variable's initial value is null as no other value has been assigned:

create variable long var_threshold

This statement creates an integer-type variable named var_output_rate and initializes it to the value ten (10):

create variable integer var_output_rate = 10

In addition to creating a variable via the create variable syntax, the runtime and engine configuration API also allows adding variables. The next code snippet illustrates the use of the runtime configuration API to create a string-typed variable:

epService.getEPAdministrator().getConfiguration()
  .addVariable("myVar", String.class, "init value");

The engine removes the variable if the statement that created the variable is destroyed and all statements that reference the variable are also destroyed. The GetVariableNameUsedBy and the RemoveVariable methods, both part of the runtime ConfigurationOperations API, provide use information and can remove a variable. If the variable was added via configuration, it can only be removed via the configuration API.

4.20.2. Setting Variable Values: the On Set clause

The on set statement assigns a new value to one or more variables when a triggering event arrives or a triggering pattern occurs. Use the SetVariableValue methods on EPRuntime to assign variable values programmatically.

The synopsis for setting variable values is:

on event_type[(filter_criteria)] [as stream_name]
  set variable_name = expression [, variable_name = expression [,...]]

The event_type is the name of the type of events that trigger the variable assignments. It is optionally followed by filter_criteria which are filter expressions to apply to arriving events. The optional as keyword can be used to assign an stream name. Patterns and named windows can also be specified in the on clause.

The comma-separated list of variable names and expressions set the value of one or more variables. Subqueries may by part of expressions however aggregation functions and the prev or prior function may not be used in expressions.

All new variable values are applied atomically: the changes to variable values by the on set statement become visible to other statements all at the same time. No changes are visible to other processing threads until the on set statement completed processing, and at that time all changes become visible at once.

The EPStatement object provides access to variable values. The pull API methods GetEnumerator and GetSafeEnumerator return the current variable values for each of the variables set by the statement. Listeners to the statement subscribe to changes in variable values: the engine posts new variable values of all variables to any listeners.

In the following example, a variable by name var_output_rate has been declared previously. When a NewOutputRateEvent event arrives, the variable is updated to a new value supplied by the event property 'rate':

on NewOutputRateEvent set var_output_rate = rate

The next example shows two variables that are updated when a ThresholdUpdateEvent arrives:

on ThresholdUpdateEvent as t 
  set var_threshold_lower = t.lower,
      var_threshold_higher = t.higher

The sample statement shown next counts the number of pattern matches using a variable. The pattern looks for OrderEvent events that are followed by CancelEvent events for the same order id within 10 seconds of the OrderEvent:

on pattern[every a=OrderEvent -> (CancelEvent(orderId=a.orderId) where timer:within(10 sec))]
  set var_counter = var_counter + 1

4.20.3. Using Variables

A variable name can be used in any expression and can also occur in an output rate limiting clause. This section presents examples and discusses performance, consistency and atomicity attributes of variables.

The next statement assumes that a variable named 'var_threshold' was created to hold a total price threshold value. The statement outputs an event when the total price for a symbol is greater then the current threshold value:

select symbol, sum(price) from TickEvent 
group by symbol 
having sum(price) > var_threshold

In this example we use a variable to dynamically change the output rate on-the-fly. The variable 'var_output_rate' holds the current rate at which the statement posts a current count to listeners:

select count(*) from TickEvent output every var_output_rate seconds

Variables are optimized towards high read frequency and lower write frequency. Variable reads do not incur locking overhead (99% of the time) while variable writes do incur locking overhead.

The engine softly guarantees consistency and atomicity of variables when your statement executes in response to an event or timer invocation. Variables acquire a stable value (implemented by versioning) when your statement starts executing in response to an event or timer invocation, and variables do not change value during execution. When one or more variable values are updated via on set statements, the changes to all updated variables become visible to statements as one unit and only when the on set statement completes successfully.

The atomicity and consistency guarantee is a soft guarantee. If any of your application statements, in response to an event or timer invocation, execute for a time interval longer then 15 seconds (default interval length), then the engine may use current variable values after 15 seconds passed, rather then then-current variable values at the time the statement started executing in response to an event or timer invocation.

The length of the time interval that variable values are held stable for the duration of execution of a given statement is by default 15 seconds, but can be configured via engine default settings.

4.20.4. Object-Type Variables

A variable of type object (or System.Object via the API) can be assigned any value including null. When using an object-type variable in an expression, your statement may need to cast the value to the desired type.

The following sample EPL creates a variable by name varobj of type object:

create variable object varobj

4.20.5. Class and Event-Type Variables

The create variable syntax and the API accept a fully-qualified class name or alternatively the name of an event type. This is useful when you want a single variable to have multiple property values to read or set.

The next statement assumes that the event type PageHitEvent is declared:

create variable PageHitEvent varPageHitZero

These example statements show two ways of assigning to the variable:

// You may assign the complete event
on PageHitEvent(ip='0.0.0.0') pagehit set varPageHitZero = pagehit
// Or assign individual properties of the event
on PageHitEvent(ip='0.0.0.0') pagehit set varPageHitZero.userId = pagehit.userId

Similarly statements may use properties of class or event-type variables as this example shows:

select * from FirewallEvent(userId=varPageHitZero.userId)

When using class or event-type variables, in order for the engine to assign property values, the underlying event type must allow writing property values. If using native event classes the class must have setter methods and a default constructor.

4.21. Contained-Event Selection

Contained-event selection is for use when an event contains properties that are themselves events. For example when application events are coarse-grained structures and you need to perform bulk operations on the rows of the property graph in an event.

Use the contained-event selection syntax in a filter expression such as in a pattern, from clause, subselect, on-select and on-delete. This section provides the synopsis and examples.

To review, in the from clause a contained_selection may appear after the event stream name and filter criteria, and before any view specifications.

The synopsis for contained_selection is as follows:

[select select_expressions from] property_expression [as property_alias] [where filter_expression]

The select clause and select_expressions are optional and may be used to select specific properties.

The property_expression is required and must be a valid property name that returns an event fragment, i.e. a property that can itself be represented as an event by the underlying event representation. Simple values such as integer or string are not fragments.

The property_alias can be provided to assign a name to the property expression.

The where clause and filter_expression is optional and may be used to filter out properties.

As an example event, consider a media order. A media order consists of order items as well as product descriptions. A media order event can be represented as an object graph (event representation), or a structure of nested Maps (Map event representation) or a XML document (XML DOM) or other custom plug-in event representation.

To illustrate, a sample media order event in XML event representation is shown below. Also, a XML event type can optionally be strongly-typed with an explicit XML XSD schema that we don't show here. Note that Map and native object representation can be considered equivalent for the purpose of this example.

Let us now assume that we have declared the event type MediaOrder as being represented by the root node <mediaorder> of such XML snip:

<mediaorder>
  <orderId>PO200901</orderId>
  <items>
    <item>
      <itemId>100001</itemId>
      <productId>B001</productId>
      <amount>10</amount>
      <price>11.95</price>
    </item>
  </items>
  <books>
    <book>
      <bookId>B001</bookId>
      <author>Heinlein</author>
      <review>
        <reviewId>1</reviewId>
        <comment>best book ever</comment>
      </review>
    </book>
    <book>
      <bookId>B002</bookId>
      <author>Isaac Asimov</author>
    </book>
  </books>
</mediaorder>

The next query utilizes the contained-event selection syntax to return each book:

select * from MediaOrder[books.book]

The result of the above query is one event per book. Output events contain only the book properties and not any of the mediaorder-level properties.

Note that, when using listeners, the engine delivers multiple results in one invocation of each listener. Therefore listeners to the above statement can expect a single invocation passing all book events within one media order event as an array.

To better illustrate the position of the contained-event selection syntax in a statement, consider the next two queries:

select * from MediaOrder(orderId='PO200901')[books.book]

The above query the returns each book only for media orders with a given order id. This query illustrates a contained-event selection and a view:

select count(*) from MediaOrder[books.book].std:unique(bookId)

The sample above counts each book unique by book id.

Contained-event selection can be staggered. When staggering multiple contained-event selections the staggered contained-event selection is relative to its parent.

This example demonstrates staggering contained-event selections by selecting each review of each book:

select * from MediaOrder[books.book][review]

Listeners to the query above receive a row for each review of each book. Output events contain only the review properties and not the book or media order properties.

The following is not valid:

// not valid
select * from MediaOrder[books.book.review]

The book property in an indexed property (an array or collection) and thereby requires an index in order to determine which book to use. The expression books.book[1].review is valid and means all reviews of the second (index 1) book.

The contained-event selection syntax is part of the filter expression and may therefore occur in patterns and anywhere a filter expression is valid.

A pattern example is below. The example assumes that a Cancel event type has been defined that also has an orderId property:

select * from pattern [c=Cancel -> books=MediaOrder(orderId = c.orderId)[books.book] ]

When used in a pattern, a filter with a contained-event selection returns an array of events, similar to the match-until clause in patterns. The above statement returns, in the books property, an array of book events.

4.21.1. Select Clause in a Contained-Event Selection

The optional select clause provides control over which fields are available in output events. The expressions in the select clause apply only to the properties available underneath the property in the from clause, and the properties of the enclosing event.

When no select is specified, only the properties underneath the selected property are available in output events.

In summary, the select clause may contain:

  1. Any expressions, wherein properties are resolved relative to the property in the from clause.

  2. Use the wildcard (*) to provide all properties that exist under the property in the from clause.

  3. Use the property_alias.* syntax to provide all properties that exist under a property in the from clause.

The next query's select clause selects each review for each book, and the order id as well as the book id of each book:

select * from MediaOrder[select orderId, bookId from books.book][select * from review]
// ... equivalent to ...
select * from MediaOrder[select orderId, bookId from books.book][review]]

Listeners to the statement above receive an event for each review of each book. Each output event has all properties of the review row, and in addition the bookId of each book and the orderId of the order. Thus bookId and orderId are found in each result event, duplicated when there are multiple reviews per book and order.

The above query uses wildcard (*) to select all properties from reviews. As has been discussed as part of the select clause, the wildcard (*) and property_alias.* do not copy properties for performance reasons. The wildcard syntax instead specifies the underlying type, and additional properties are added onto that underlying type if required. Only one wildcard (*) and property_alias.* (unless used with a column rename) may therefore occur in the select clause list of expressions.

All the following queries produce an output event for each review of each book. The next sample queries illustrate the options available to control the fields of output events.

The output events produced by the next query have all properties of each review and no other properties available:

select * from MediaOrder[books.book][review]

The following query is not a valid query, since the order id and book id are not part of the contained-event selection:

// Invalid select clause: orderId and bookId not produced.
select orderId, bookId from MediaOrder[books.book][review]

This query is valid. Note that output events carry only the orderId and bookId properties and no other data:

select orderId, bookId from MediaOrder[books.book][select orderId, bookId from review]
//... equivalent to ...
select * from MediaOrder[select orderId, bookId from books.book][review]

This variation produces output events that have all properties of each book and only reviewId and comment for each review:

select * from MediaOrder[select * from books.book][select reviewId, comment from review]
// ... equivalent to ...
select * from MediaOrder[books.book as book][select book.*, reviewId, comment from review]

The output events of the next EPL have all properties of the order and only bookId and reviewId for each review:

select * from MediaOrder[books.book as book]
    [select mediaOrder.*, bookId, reviewId from review] as mediaOrder

This EPL produces output events with 3 columns: a column named mediaOrder that is the order itself, a column named book for each book and a column named review that holds each review:

insert into ReviewStream select * from MediaOrder[books.book as book]
  [select mo.* as mediaOrder, book.* as book, review.* as review from review as review] as mo
// .. and a sample consumer of ReviewStream...
select mediaOrder.orderId, book.bookId, review.reviewId from ReviewStream

Please note these limitations:

  1. Sub-selects, aggregation functions and the prev and prior operators are not available in contained-event selection.

  2. Expressions in the select and where clause of a contained-event selection can only reference properties relative to the current event and property.

4.21.2. Where Clause in a Contained-Event Selection

The optional where clause may be used to filter out properties at the same level that the where-clause occurs.

The properties in the filter expression must be relative to the property in the from clause or the enclosing event.

This query outputs all books with a given author:

select * from MediaOrder[books.book where author = 'Heinlein']

This query outputs each review of each book where a review comment contains the word 'good':

select * from MediaOrder[books.book][review where comment like 'good']

4.21.3. Contained-Event Selection and Joins

This section discusses contained-event selection in joins.

When joining within the same event it is not required that views are specified. Recall, in a join or outer join there must be views specified that hold the data to be joined. For self-joins, no views are required and the join executes against the data returned by the same event.

This query inner-joins items to books where book id matches the product id:

select book.bookId, item.itemId 
from MediaOrder[books.book] as book, 
      MediaOrder[items.item] as item 
where productId = bookId

Query results for the above query when sending the media order event as shown earlier are:

book.bookIditem.itemId
B001100001

The next example query is a left outer join. It returns all books and their items, and for books without item it returns the book and a null value:

select book.bookId, item.itemId 
from MediaOrder[books.book] as book 
  left outer join 
    MediaOrder[items.item] as item 
  on productId = bookId

Query results for the above query when sending the media order event as shown earlier are:

book.bookIditem.itemId
B001100001
B002null

A full outer join combines the results of both left and right outer joins. The joined table will contain all records from both tables, and fill in null values for missing matches on either side.

This example query is a full outer join, returning all books as well as all items, and filling in null values for book id or item id if no match is found:

select orderId, book.bookId,item.itemId 
from MediaOrder[books.book] as book 
  full outer join 
     MediaOrder[select orderId, * from items.item] as item 
  on productId = bookId 
order by bookId, item.itemId asc

As in all other continuous queries, aggregation results are cumulative from the time the statement was created.

The following query counts the cumulative number of items in which the product id matches a book id:

select count(*) 
from MediaOrder[books.book] as book, 
      MediaOrder[items.item] as item 
where productId = bookId

The unidirectional keyword in a join indicates to the query engine that aggregation state is not cumulative. The next query counts the number of items in which the product id matches a book id for each event:

select count(*) 
from MediaOrder[books.book] as book unidirectional, 
      MediaOrder[items.item] as item 
where productId = bookId

4.22. Updating an Insert Stream: the Update IStream Clause

The update istream statement allows declarative modification of event properties of events entering a stream. Update is a pre-processing step to each new event, modifying an event before the event applies to any statements.

The synopsis of update istream is as follows:

update istream event_type [as stream_name]
  set property_name = set_expression [, property_name = set_expression] [,...]
  [where where_expression]

The event_type is the name of the type of events that the update applies to. The optional as keyword can be used to assign a name to the event type for use with subqueries, for example. Following the set keyword is a comma-separated list of property names and expressions that provide the event properties to change and values to set.

The optional where clause and expression can be used to filter out events to which to apply updates.

Listeners to an update statement receive the updated event in the insert stream (new data) and the event prior to the update in the remove stream (old data). Note that if there are multiple update statements that all apply to the same event then the engine will ensure that the output events delivered to listeners or subscribers are consistent with the then-current updated properties of the event (if necessary making event copies, as described below, in the case that listeners are attached to update statements). Iterating over an update statement returns no events.

As an example, the below statement assumes an AlertEvent event type that has properties named severity and reason:

update istream AlertEvent 
  set severity = 'High'
  where severity = 'Medium' and reason like '%withdrawal limit%'

The statement above changes the value of the severity property to "High" for AlertEvent events that have a medium severity and contain a specific reason text.

Update statements apply the changes to event properties before other statements receive the event(s) for processing, e.g. "select * from AlertEvent" receives the updated AlertEvent. This is true regardless of the order in which your application creates statements.

When multiple update statements apply to the same event, the engine executes updates in the order in which update statements are created. We recommend the @Priority EPL annotation to define a deterministic order of processing updates, especially in the case where update statements get created and destroyed dynamically or multiple update statements update the same fields. The update statement with the highest @Priority value applies last.

The update clause can be used on streams populated via insert into, as this example utilizing a pattern demonstrates:

insert into DoubleWithdrawalStream 
select a.id, b.id, a.account as account, 0 as minimum 
from pattern [a=Withdrawal -> b=Withdrawal(id = a.id)]
update istream DoubleWithdrawalStream set minimum = 1000 where account in (10002, 10003)

When using update with named windows, any changes to event properties apply before an event enters the named window.

Consider the next example (shown here with statement names in @Name EPL annotation, multiple EPL statements):

@Name("CreateWindow") create window MyWindow.win:time(30 sec) as AlertEvent

@Name("UpdateStream") update istream MyWindow set severity = 'Low' where reason = '%out of paper%'

@Name("InsertWindow") insert into MyWindow select * from AlertEvent

@Name("SelectWindow") select * from MyWindow

The UpdateStream statement specifies an update clause that applies to all events entering the named window. Note that update does not apply to events already in the named window at the time an application creates the UpdateStream statement, it only applies to new events entering the named window (after an application created the update statement).

Therefore, in the above example listeners to the SelectWindow statement as well as the CreateWindow statement receive the updated event, while listeners to the InsertWindow statement receive the original AlertEvent event (and not the updated event).

Subqueries can also be used in all expressions including the optional where clause.

This example demonstrates a correlated subquery in an assignment expression and also demonstrates the optional as keyword. It assigns the phone property of an AlertEvent event a new value based on the lookup within all unique PhoneEvent events (according to an empid property) correlating the AlertEvent property reporter with the empid property of PhoneEvent:

update istream AlertEvent as ae
  set phone = 
    (select phone from PhoneEvent.std:unique(empid) where empid = ae.reporter)

When using update, please note these limitations:

  1. Expressions may not use aggregation functions.

  2. The prev and prior functions may not be used.

  3. For underlying event representations that are CLR objects, a event object class must implement the Serializable behavior as discussed below.

  4. When using an XML underlying event type, event properties in the XML document representation are not available for update.

  5. Nested, indexed and mapped properties are not supported for update. Revision event types and variant streams may also not be updated.

4.22.1. Immutability and Updates

When updating event objects the engine maintains consistency across statements. The engine ensures that an update to an event does not impact the results of statements that look for or retain the original un-updated event. As a result the engine may need to copy an event object to maintain consistency.

In the case your application utilizes objects as the underlying event representation and an update statement updates properties on an object, then in order to maintain consistency across statements it is necessary for the engine to copy the object before changing properties (and thus not change the original object).

For application objects, the copy operation is implemented by serialization. Your event object must therefore implement the Serializable behavior to become eligible for update. As an alternative to serialization, you may instead configure a copy method as part of the event type configuration via ConfigurationEventTypeLegacy.

4.23. Controlling Event Delivery : The For Clause

The engine delivers all result events of a given statement to the statement's listeners and subscriber (if any) in a single invocation of each listener and subscriber's Update method passing an array of result events. For example, a statement using a time-batch view may provide many result events after a time period passes, a pattern may provide multiple matching events or in a join the join cardinality could be multiple rows.

For statements that typically post multiple result events to listeners the for keyword controls the number of invocations of the engine to listeners and subscribers and the subset of all result events delivered by each invocation. This can be useful when your application listener or subscriber code expects multiple invocations or expects that invocations only receive events that belong together by some additional criteria.

The for keyword is a reserved keyword. It is followed by either the grouped_delivery keyword for grouped delivery or the discrete_delivery keyword for discrete delivery. The for clause is valid after any EPL select statement.

The synopsis for grouped delivery is as follows:

... for grouped_delivery (group_expression [, group_expression] [,...])

The group_expression expression list provides one or more expressions to apply to result events. The engine invokes listeners and subscribers once for each distinct set of values returned by group_expression expressions passing only the events for that group.

The synopsis for discrete delivery is as follows:

... for discrete_delivery

With discrete delivery the engine invokes listeners and subscribers once for each result event passing a single result event in each invocation.

Consider the following example without for-clause. The time batch data view collects RFIDEvent events for 10 seconds and posts an array of result events:

select * from RFIDEvent.win:time_batch(10 sec)

Let's consider an example event sequence as follows:

Table 4.4. Sample Sequence of Events for For Keyword

Event
RFIDEvent(id:1, zone:'A')
RFIDEvent(id:2, zone:'B')
RFIDEvent(id:3, zone:'A')

Without for-clause and after the 10-second time period passes, the engine delivers an array of 3 events in a single invocation to listeners and the subscriber.

The next example specifies the for-clause and grouped delivery by zone:

select * from RFIDEvent.win:time_batch(10 sec) for grouped_delivery (zone)

With grouped delivery and after the 10-second time period passes, the above statement delivers result events in two invocations to listeners and the subscriber: The first invocation delivers an array of two events that contains zone A events with id 1 and 3. The second invocation delivers an array of 1 event that contains a zone B event with id 2.

The next example specifies the for-clause and discrete delivery:

select * from RFIDEvent.win:time_batch(10 sec) for discrete_delivery

With discrete delivery and after the 10-second time period passes, the above statement delivers result events in three invocations to listeners and the subscriber: The first invocation delivers an array of 1 event that contains the event with id 1, the second invocation delivers an array of 1 event that contains the event with id 2 and the third invocation delivers an array of 1 event that contains the event with id 3.

Remove stream events are also delivered in multiple invocations, one for each group, if your statement selects remove stream events explicitly via irstream or rstream keywords.

The insert into for inserting events into a stream is not affected by the for-clause.

The delivery order respects the natural sort order or the explicit sort order as provided by the order by clause, if present.

The following are known limitations:

  1. The engine validates group_expression expressions against the output event type, therefore all properties specified in group_expression expressions must occur in the select clause.

Chapter 5. EPL Reference: Patterns

5.1. Event Pattern Overview

Event patterns match when an event or multiple events occur that match the pattern's definition. Patterns can also be time-based.

Pattern expressions consist of pattern atoms and pattern operators:

  1. Pattern atoms are the basic building blocks of patterns. Atoms are filter expressions, observers for time-based events and plug-in custom observers that observe external events not under the control of the engine.

  2. Pattern operators control expression lifecycle and combine atoms logically or temporally.

The below table outlines the different pattern atoms available:

Table 5.1. Pattern Atoms

Pattern AtomExample
Filter expressions specify an event to look for.
StockTick(symbol='ABC', price > 100)
Time-based event observers specify time intervals or time schedules.
timer:interval(10 seconds)
timer:at(*, 16, *, *, *)
Custom plug-in observers can add pattern language syntax for observing application-specific events.
myapplication:myobserver("http://someResource")

There are 4 types of pattern operators:

  1. Operators that control pattern sub-expression repetition: every, every-distinct, [num] and until

  2. Logical operators: and, or, not

  3. Temporal operators that operate on event order: -> (followed-by)

  4. Guards are where-conditions that control the lifecycle of subexpressions. Examples are timer:within, timer:withinmax and while-expression. Custom plug-in guards may also be used.

Pattern expressions can be nested arbitrarily deep by including the nested expression(s) in () round parenthesis.

Underlying the pattern matching is a state machine that transitions between states based on arriving events and based on time advancing. A single event or advancing time may cause a reaction in multiple parts of your active pattern state.

5.2. How to use Patterns

5.2.1. Pattern Syntax

This is an example pattern expression that matches on every ServiceMeasurement events in which the value of the latency event property is over 20 seconds, and on every ServiceMeasurement event in which the success property is false. Either one or the other condition must be true for this pattern to match.

every (spike=ServiceMeasurement(latency>20000) or error=ServiceMeasurement(success=false))

In the example above, the pattern expression starts with an every operator to indicate that the pattern should fire for every matching events and not just the first matching event. Within the every operator in round brackets is a nested pattern expression using the or operator. The left hand of the or operator is a filter expression that filters for events with a high latency value. The right hand of the operator contains a filter expression that filters for events with error status. Filter expressions are explained in Section 5.4, “Filter Expressions In Patterns”.

The example above assigned the tags spike and error to the events in the pattern. The tags are important since the engine only places tagged events into the output event(s) that a pattern generates, and that the engine supplies to listeners of the pattern statement. The tags can further be selected in the select-clause of an EPL statement as discussed in Section 4.4.2, “Pattern-based Event Streams”.

Patterns can also contain comments within the pattern as outlined in Section 4.2.2, “Using Comments”.

Pattern statements are created via the EPAdministrator interface. The EPAdministrator interface allows to create pattern statements in two ways: Pattern statements that want to make use of the EPL select clause or any other EPL constructs use the CreateEPL method to create a statement that specifies one or more pattern expressions. EPL statements that use patterns are described in more detail in Section 4.4.2, “Pattern-based Event Streams”. Use the syntax as shown in below example.

EPAdministrator admin = EPServiceProviderManager.GetDefaultProvider().EPAdministrator;

String eventName = ServiceMeasurement.Name;

EPStatement myTrigger = admin.CreateEPL("select * from pattern [" +
  "every (spike=" + eventName + "(latency>20000) or error=" + eventName + "(success=false))]");

Pattern statements that do not need to make use of the EPL select clause or any other EPL constructs can use the CreatePattern method, as in below example.

EPStatement myTrigger = admin.CreatePattern(
  "every (spike=" + eventName + "(latency>20000) or error=" + eventName + "(success=false))");

5.2.2. Patterns in EPL

A pattern may appear anywhere in the from clause of an EPL statement including joins and subqueries. Patterns may therefore be used in combination with the where clause, group by clause, having clause as well as output rate limiting and insert into.

In addition, a data window view can be declared onto a pattern. A data window declared onto a pattern only serves to retain pattern matches. A data window declared onto a pattern does not limit, cancel, remove or delete intermediate pattern matches of the pattern when pattern matches leave the data window.

This example statement demonstrates the idea by selecting a total price per customer over pairs of events (ServiceOrder followed by a ProductOrder event for the same customer id within 1 minute), occuring in the last 2 hours, in which the sum of price is greater than 100, and using a where clause to filter on name:

select a.custId, sum(a.price + b.price)
from pattern [every a=ServiceOrder -> 
    b=ProductOrder(custId = a.custId) where timer:within(1 min)].win:time(2 hour) 
where a.name in ('Repair', b.name)
group by a.custId
having sum(a.price + b.price) > 100

5.2.3. Subscribing to Pattern Events

When a pattern fires it publishes one or more events to any event handlers attached to the pattern statement. Event handlers are required to implement the com.espertech.esper.client.UpdateEventHandler delegate.

The example below shows an implementation of the com.espertech.esper.client.UpdateEventHandler delegate. We add the anonymous delegate to the myPattern statement created earlier. The listener code simply extracts the underlying event class.

myPattern.Events += 
  delegate(Object sender, UpdateEventArgs e) {
    ServiceMeasurement spike = (ServiceMeasurement) e.NewEvents[0].Get("spike");
    ServiceMeasurement error = (ServiceMeasurement) e.NewEvents[0].Get("error");
    ... // either spike or error can be null, depending on which occurred
    ... // add more logic here
  };

Event handlers receive the sender of the event (the statement) and an event that indicates details about the event that has occurred. The event includes an array of EventBean instances in the NewEvents property. There is one EventBean instance passed to the event handler for each combination of events that matches the pattern expression. At least one EventBean instance is always passed to the event handler.

The properties of each EventBean instance contain the underlying events that caused the pattern to fire, if events have been named in the filter expression via the name=eventType syntax. The property name is thus the name supplied in the pattern expression, while the property type is the type of the underlying class, in this example ServiceMeasurement.

5.2.4. Pulling Data from Patterns

Data can also be obtained from pattern statements via the GetSafeEnumerator() and GetEnumerator() methods on EPStatement (the pull API). If the pattern had fired at least once, then the iterator returns the last event for which it fired. The MoveNext() method can be used to determine if the pattern had fired. Additionally, extension methods such as HasFirst and FirstOrDefault can be applied to the statement because it implements the IEnumerable interface.

if (myPattern.HasFirst()) {
	ServiceMeasurement event = (ServiceMeasurement) view.FirstOrDefault().Get("alert");
    ... // some more code here to process the event
}
else {
    ... // no matching events at this time
}

Further, if a data window is defined onto a pattern, the iterator returns the pattern matches according to the data window expiry policy.

This pattern specifies a length window of 10 elements that retains the last 10 matches of A and B events, for use via iterator or for use in a join or subquery:

select * from pattern [every (A or B).win:length(10)

5.2.5. Pattern Error Reporting

While the pattern compiler analyzes your pattern and verifies its integrity, it may not detect certain pattern errors that may occur at runtime. Sections of this pattern documentation point out common cases where the pattern engine will log a runtime error. We recommend turning on the log warning level at project development time to inspect and report on warnings logged. If a statement name is assigned to a statement then the statement name is logged as well.

5.3. Operator Precedence

The operators at the top of this table take precedence over operators lower on the table.

Table 5.2. Pattern Operator Precedence

PrecedenceOperatorDescriptionExample
1unaryevery, not, every-distinct
every MyEvent
timer:interval(5 min) and not MyEvent
2guard postfixwhere timer:within and while (expression) (incl. withinmax and plug-in pattern guard)
MyEvent where timer:within(1 sec)
a=MyEvent while (a.price between 1 and 10)
3repeat[num], until
[5] MyEvent
[1..3] MyEvent until MyOtherEvent
4andand
every (MyEvent and MyOtherEvent)
5oror
every (MyEvent or MyOtherEvent)
6followed-by->
every (MyEvent -> MyOtherEvent)

If you are not sure about the precedence, please consider placing parenthesis () around your subexpressions. Parenthesis can also help make expressions easier to read and understand.

Note that we are also providing the EPL grammar as a HTML file as part of the documentation set on the project website.

The following table outlines sample equivalent expressions, with and without the use of parenthesis for subexpressions.

Table 5.3. Equivalent Pattern Expressions

ExpressionEquivalentReason
every A or B(every A) or BThe every operator has higher precedence then the or operator.
every A -> B or C(every A) -> (B or C)The or operator has higher precedence then the followed-by operator.
A -> B or B -> AA -> (B or B) -> AThe or operator has higher precedence then the followed-by operator, specify as (A -> B) or (B -> A) instead.
A and B or C(A and B) or CThe and operator has higher precedence then the or operator.
every A where timer:within(5)every (A where timer:within(5))The every operator has higher precedence then the timer:within guard postfix.
A -> B until C -> DA -> (B until C) -> DThe until operator has higher precedence then the followed-by operator.
[5] A or B ([5] A) or BThe [num] repeat operator has higher precedence then the or operator.

5.4. Filter Expressions In Patterns

The simplest form of filter is a filter for events of a given type without any conditions on the event property values. This filter matches any event of that type regardless of the event's properties. The example below is such a filter. Note that this event pattern would stop firing as soon as the first RfidEvent is encountered.

com.mypackage.myevents.RfidEvent

To make the event pattern fire for every RfidEvent and not just the first event, use the every keyword.

every com.mypackage.myevents.RfidEvent

The example above specifies the fully-qualified type name as the event type. Via configuration, the event pattern above can be simplified by using the name that has been defined for the event type.

every RfidEvent

Interfaces and superclasses are also supported as event types. In the below example IRfidReadable is an interface class, and the statement matches any event that implements this interface:

every org.myorg.rfid.IRfidReadable

The filtering criteria to filter for events with certain event property values are placed within parenthesis after the event type name:

RfidEvent(category="Perishable")

All expressions can be used in filters, including static method invocations that return a boolean value:

RfidEvent(com.mycompany.MyRFIDLib.isInRange(x, y) or (x<0 and y < 0))

Filter expressions can be separated via a single comma ','. The comma represents a logical AND between expressions:

RfidEvent(zone=1, category=10)
...is equivalent to...
RfidEvent(zone=1 and category=10)

The following set of operators are highly optimized through indexing and are the preferred means of filtering high-volume event streams:

  • equals =

  • not equals !=

  • comparison operators < , > , >=, <=

  • ranges

    • use the between keyword for a closed range where both endpoints are included

    • use the in keyword and round () or square brackets [] to control how endpoints are included

    • for inverted ranges use the not keyword and the between or in keywords

  • list-of-values checks using the in keyword or the not in keywords followed by a comma-separated list of values

At compile time as well as at run time, the engine scans new filter expressions for subexpressions that can be indexed. Indexing filter values to match event properties of incoming events enables the engine to match incoming events faster. The above list of operators represents the set of operators that the engine can best convert into indexes. The use of comma or logical and in filter expressions does not impact optimizations by the engine.

For more information on filters please see Section 4.4.1, “Filter-based Event Streams”. Contained-event selection on filters in patterns is further described in Section 4.21, “Contained-Event Selection”.

Filter criteria can also refer to events matching prior named events in the same expression. Below pattern is an example in which the pattern matches once for every RfidEvent that is preceded by an RfidEvent with the same asset id.

every e1=RfidEvent -> e2=RfidEvent(assetId=e1.assetId)

The syntax shown above allows filter criteria to reference prior results by specifying the event name tag of the prior event, and the event property name. The tag names in the above example were e1 and e2. This syntax can be used in all filter operators or expressions including ranges and the in set-of-values check:

every e1=RfidEvent -> 
  e2=RfidEvent(MyLib.isInRadius(e1.x, e1.y, x, y) and zone in (1, e1.zone))

An arriving event changes the truth value of all expressions that look for the event. Consider the pattern as follows:

every (RfidEvent(zone > 1) and RfidEvent(zone < 10))

The pattern above is satisfied as soon as only one event with zone in the interval [2, 9] is received.

5.5. Pattern Operators

5.5.1. Every

The every operator indicates that the pattern sub-expression should restart when the subexpression qualified by the every keyword evaluates to true or false. Without the every operator the pattern sub-expression stops when the pattern sub-expression evaluates to true or false.

As a side note, please be aware that a single invocation to the UpdateEventHandler may deliver multiple events in one invocation, since the UpdateEventArgs contain an array of values.

Thus the every operator works like a factory for the pattern sub-expression contained within. When the pattern sub-expression within it fires and thus quits checking for events, the every causes the start of a new pattern sub-expression listening for more occurrences of the same event or set of events.

Every time a pattern sub-expression within an every operator turns true the engine starts a new active subexpression looking for more event(s) or timing conditions that match the pattern sub-expression. If the every operator is not specified for a subexpression, the subexpression stops after the first match was found.

This pattern fires when encountering an A event and then stops looking.

A

This pattern keeps firing when encountering A events, and doesn't stop looking.

every A

When using every operator with the -> followed-by operator, each time the every operator restarts it also starts a new subexpression instance looking for events in the followed-by subexpression.

Let's consider an example event sequence as follows.

A1   B1   C1   B2   A2   D1   A3   B3   E1   A4   F1   B4

Table 5.4. 'Every' operator examples

ExampleDescription
every ( A -> B )

Detect an A event followed by a B event. At the time when B occurs the pattern matches, then the pattern matcher restarts and looks for the next A event.

  1. Matches on B1 for combination {A1, B1}

  2. Matches on B3 for combination {A2, B3}

  3. Matches on B4 for combination {A4, B4}

every A -> B

The pattern fires for every A event followed by a B event.

  1. Matches on B1 for combination {A1, B1}

  2. Matches on B3 for combination {A2, B3} and {A3, B3}

  3. Matches on B4 for combination {A4, B4}

A -> every B

The pattern fires for an A event followed by every B event.

  1. Matches on B1 for combination {A1, B1}.

  2. Matches on B2 for combination {A1, B2}.

  3. Matches on B3 for combination {A1, B3}

  4. Matches on B4 for combination {A1, B4}

every A -> every B

The pattern fires for every A event followed by every B event.

  1. Matches on B1 for combination {A1, B1}.

  2. Matches on B2 for combination {A1, B2}.

  3. Matches on B3 for combination {A1, B3} and {A2, B3} and {A3, B3}

  4. Matches on B4 for combination {A1, B4} and {A2, B4} and {A3, B4} and {A4, B4}

The examples show that it is possible that a pattern fires for multiple combinations of events that match a pattern expression. Each combination is posted as an EventBean instance to the UpdateEventHandler implementation.

Let's consider the every operator in conjunction with a subexpression that matches 3 events that follow each other:

every (A -> B -> C)

The pattern first looks for A events. When an A event arrives, it looks for a B event. After the B event arrives, the pattern looks for a C event. Finally, when the C event arrives the pattern fires. The engine then starts looking for an A event again.

Assume that between the B event and the C event a second A2 event arrives. The pattern would ignore the A2 event entirely since it's then looking for a C event. As observed in the prior example, the every operator restarts the subexpression A -> B -> C only when the subexpression fires.

In the next statement the every operator applies only to the A event, not the whole subexpression:

every A -> B -> C

This pattern now matches for each A event that is followed by a B event and then a C event, regardless of when the A event arrives. Note that for each A event that arrives the pattern engine starts a new subexpression looking for a B event and then a C event, outputting each combination of matching events.

5.5.1.1. Limiting Subexpression Lifetime

As the introduction of the every operator states, the operator starts new subexpression instances and can cause multiple matches to occur for a single arriving event.

New subexpressions also take a very small amount of system resources and thereby your application should carefully consider when subexpressions must end when designing patterns. Use the timer:within construct and the and not constructs to end active subexpressions. The data window onto a pattern stream does not serve to limit pattern sub-expression lifetime.

Lets look at a concrete example. Consider the following sequence of events arriving:

A1   A2   B1  

This pattern matches on arrival of B1 and outputs two events (an array of length 2 if using a listener). The two events are the combinations {A1, B1} and {A2, B1}:

every a=A -> b=B

The and not operators are used to end an active subexpression.

The next pattern matches on arrival of B1 and outputs only the last A event which is the combination {A2, B1}:

every a=A -> (b=B and not A)

The and not operators cause the subexpression looking for {A1, B?} to end when A2 arrives.

Similarly, in the pattern below the engine starts a new subexpression looking for a B event every 1 second. After 5 seconds there are 5 subexpressions active looking for a B event and 5 matches occur at once if a B event arrives after 5 seconds.

every timer:interval(1 sec) -> b=B

Again the and not operators can end subexpressions that are not intended to match any longer:

every timer:interval(1 sec) -> (b=B and not timer:interval(1 sec))
// equivalent to
every timer:interval(1 sec) -> (b=B where timer:within(1 sec))

5.5.1.2. Every Operator Example

In this example we consider a generic pattern in which the pattern must match for each A event followed by a B event and followed by a C event, in which both the B event and the C event must arrive within 1 hour of the A event. The first approach to the pattern is as follows:

every A  -> (B -> C) where timer:within(1 hour)

Consider the following sequence of events arriving:

A1   A2   B1   C1   B2   C2

First, the pattern as above never stops looking for A events since the every operator instructs the pattern to keep looking for A events.

When A1 arrives, the pattern starts a new subexpression that keeps A1 in memory and looks for any B event. At the same time, it also keeps looking for more A events.

When A2 arrives, the pattern starts a new subexpression that keeps A2 in memory and looks for any B event. At the same time, it also keeps looking for more A events.

After the arrival of A2, there are 3 subexpressions active:

  1. The first active subexpression with A1 in memory, looking for any B event.

  2. The second active subexpression with A2 in memory, looking for any B event.

  3. A third active subexpression, looking for the next A event.

In the pattern above, we have specified a 1-hour lifetime for subexpressions looking for B and C events. Thus, if no B and no C event arrive within 1 hour after A1, the first subexpression goes away. If no B and no C event arrive within 1 hour after A2, the second subexpression goes away. The third subexpression however stays around looking for more A events.

The pattern as shown above thus matches on arrival of C1 for combination {A1, B1, C1} and for combination {A2, B1, C1}, provided that B1 and C1 arrive within an hour of A1 and A2.

You may now ask how to match on {A1, B1, C1} and {A2, B2, C2} instead, since you may need to correlate on a given property.

The pattern as discussed above matches every A event followed by the first B event followed by the next C event, and doesn't specifically qualify the B or C events to look for based on the A event. To look for specific B and C events in relation to a given A event, the correlation must use one or more of the properties of the A event, such as the "id" property:

every a=A -> (B(id=a.id -> C(id=a.id)) where timer:within(1 hour)

The pattern as shown above thus matches on arrival of C1 for combination {A1, B1, C1} and on arrival of C2 for combination {A2, B2, C2}.

5.5.1.3. Sensor Example

This example looks at temperature sensor events named Sample. The pattern detects when 3 sensor events indicate a temperature of more then 50 degrees uninterrupted within 90 seconds of the first event, considering events for the same sensor only.

every sample=Sample(temp > 50) ->
( (Sample(sensor=sample.sensor, temp > 50) and not Sample(sensor=sample.sensor, temp <= 50))   
  ->
  (Sample(sensor=sample.sensor, temp > 50) and not Sample(sensor=sample.sensor, temp <= 50))   
 ) where timer:within(90 seconds))

The pattern starts a new subexpression in the round braces after the first followed-by operator for each time a sensor indicated more then 50 degrees. Each subexpression then lives a maximum of 90 seconds. Each subexpression ends if a temperature of 50 degress or less is encountered for the same sensor. Only if 3 temperature events in a row indicate more then 50 degrees, and within 90 seconds of the first event, and for the same sensor, does this pattern fire.

5.5.2. Every-Distinct

Similar to the every operator in most aspects, the every-distinct operator indicates that the pattern sub-expression should restart when the subexpression qualified by the every-distinct keyword evaluates to true or false. In addition, the every-distinct eliminates duplicate results received from an active subexpression according to its distinct-value expressions.

The synopsis for the every-distinct pattern operator is:

every-distinct(distinct_value_expr [, distinct_value_exp[...][, expiry_time_period])

Within parenthesis are one or more distinct_value_expr expressions that return the values by which to remove duplicates.

You may optionally specify an expiry_time_period time period. If present, the pattern engine expires and removes distinct key values that are older then the time period, removing their associated memory and allowing such distinct values to match again. When your distinct value expressions return an unlimited number of values, for example when your distinct value is a timestamp or auto-increment column, you should always specify an expiry time period.

When specifying properties in the distinct-value expression list, you must ensure that the event types providing properties are tagged. Only properties of event types within filter expressions that are sub-expressions to the every-distinct may be specified.

For example, this pattern keeps firing for every A event with a distinct value for its aprop property:

every-distinct(a.aprop) a=A

Note that the pattern above assigns the a tag to the A event and uses a.prop to identify the prop property as a value of the a event A.

A pattern that returns the first Sample event for each sensor, assuming sensor is a field that returns a unique id identifying the sensor that originated the Sample event, is:

every-distinct(s.sensor) s=Sample

The next pattern looks for pairs of A and B events and returns only the first pair for each combination of aprop of an A event and bprop of a B event:

every-distinct(a.aprop, b.bprop) (a=A and b=B)

The following pattern looks for A events followed by B events for which the value of the aprop of an A event is the same value of the bprop of a B event but only for each distinct value of aprop of an A event:

every-distinct(a.aprop) a=A -> b=B(bprop = a.aprop)

When specifying properties as part of distinct-value expressions, properties must be available from tagged event types in sub-expressions to the every-distinct.

The following patterns are not valid:

// Invalid: event type in filter not tagged
every-distinct(aprop) A
			
// Invalid: property not from a sub-expression of every-distinct
a=A -> every-distinct(a.aprop) b=B

When an active subexpression to every-distinct becomes permanently false, the distinct-values seen from the active subexpression are removed and the sub-expression within is restarted.

For example, the below pattern detects each A event distinct by the value of aprop.

every-distinct(a.aprop) (a=A and not B)

In the pattern above, when a B event arrives, the subexpression becomes permanently false and is restarted anew, detecting each A event distinct by the value of aprop without considering prior values.

When your distinct key is a timestamp or other non-unique property, specify an expiry time period.

The following example returns every distinct A event according to the timestamp property on the A event, retaining each timestamp value for 10 seconds:

every-distinct(a.timestamp, 10 seconds) a=A

In the example above, if for a given A event and its timestamp value the same timestamp value occurs again for another A event before 10 seconds passed, the A event is not a match. If 10 seconds passed the pattern indicates a second match.

5.5.3. Repeat

The repeat operator fires when a pattern sub-expression evaluates to true a given number of times. The synopsis is as follows:

[match_count] repeating_subexpr

The repeat operator is very similar to the every operator in that it restarts the repeating_subexpr pattern sub-expression up to a given number of times.

match_count is a positive number that specifies how often the repeating_subexpr pattern sub-expression must evaluate to true before the repeat expression itself evaluates to true, after which the engine may indicate a match.

For example, this pattern fires when the last of five A events arrives:

[5] A

Parenthesis must be used for nested pattern sub-expressions. This pattern fires when the last of a total of any five A or B events arrives:

[5] (A or B)

Without parenthesis the pattern semantics change, according to the operator precedence described earlier. This pattern fires when the last of a total of five A events arrives or a single B event arrives, whichever happens first:

[5] A or B

Tags can be used to name events in filter expression of pattern sub-expressions. The next pattern looks for an A event followed by a B event, and a second A event followed by a second B event. The output event provides indexed and array properties of the same name:

[2] (a=A -> b=B)

Using tags with repeat is further described in Section 5.5.4.6, “Tags and the Repeat Operator”.

Consider the following pattern that demonstrates the behavior when a pattern sub-expression becomes permanently false:

[2] (a=A and not C)

In the case where a C event arrives before 2 A events arrive, the pattern above becomes permanently false.

Lets add an every operator to restart the pattern and thus keep matching for all pairs of A events that arrive without a C event in between each pair:

every [2] (a=A and not C)

Since pattern matches return multiple A events, your select clause should use tag a as an array, for example:

select a[0].id, a[1].id from pattern [every [2] (a=A and not C)]

5.5.4. Repeat-Until

The repeat until operator provides additional control over repeated matching.

The repeat until operator takes an optional range, a pattern sub-expression to repeat, the until keyword and a second pattern sub-expression that ends the repetition. The synopsis is as follows:

[range] repeated_pattern_expr until end_pattern_expr

Without a range, the engine matches the repeated_pattern_expr pattern sub-expression until the end_pattern_expr evaluates to true, at which time the expression turns true.

An optional range can be used to indicate the minimum number of times that the repeated_pattern_expr pattern sub-expression must become true.

The optional range can also specify a maximum number of times that repeated_pattern_expr pattern sub-expression evaluates to true and retains tagged events. When this number is reached, the engine stops the repeated_pattern_expr pattern sub-expression.

5.5.4.1. Unbound Repeat

In the unbound repeat, without a range, the engine matches the repeated_pattern_expr pattern sub-expression until the end_pattern_expr evaluates to true, at which time the expression turns true. The synopsis is:

repeated_pattern_expr until end_pattern_expr

This is a pattern that keeps looking for A events until a B event arrives:

A until B

Nested pattern sub-expressions must be placed in parenthesis since the until operator has precedence over most operators. This example collects all A or B events for 10 seconds and places events received in indexed properties 'a' and 'b':

(a=A or b=B) until timer:interval(10 sec)

5.5.4.2. Bound Repeat Overview

The synopsis for the optional range qualifier is:

[ [low_endpoint] : [high_endpoint] ]

low_endpoint is an optional number that appears on the left of a colon (:), after which follows an optional high_endpoint number.

A range thus consists of a low_endpoint and a high_endpoint in square brackets and separated by a colon (:) characters. Both endpoint values are optional but either one or both must be supplied. The low_endpoint can be omitted to denote a range that starts at zero. The high_endpoint can be omitted to denote an open-ended range.

Some examples for valid ranges might be:

[3 : 10]
[:3]    // range starts at zero
[2:]    // open-ended range

The low_endpoint, if specified, defines the minimum number of times that the repeated_pattern_expr pattern sub-expression must become true in order for the expression to become true.

The high_endpoint, if specified, is the maximum number of times that the repeated_pattern_expr pattern sub-expression becomes true. If the number is reached, the engine stops the repeated_pattern_expr pattern sub-expression.

In all cases, only at the time that the end_pattern_expr pattern sub-expression evaluates to true does the expression become true. If end_pattern_expr pattern sub-expression evaluates to false, then the expression evaluates to false.

5.5.4.3. Bound Repeat - Open Ended Range

An open-ended range specifies only a low endpoint and not a high endpoint.

Consider the following pattern which requires at least three A events to match:

[3:] A until B

In the pattern above, if a B event arrives before 3 A events occurred, the expression ends and evaluates to false.

5.5.4.4. Bound Repeat - High Endpoint Range

A high-endpoint range specifies only a high endpoint and not a low endpoint.

In this sample pattern the engine will be looking for a maximum of 3 A events. The expression turns true as soon as a single B event arrives regardless of the number of A events received:

[:3] A until B

The next pattern matches when a C or D event arrives, regardless of the number of A or B events that occurred:

[:3] (a=A or b=B) until (c=C or d=D)

In the pattern above, if more then 3 A or B events arrive, the pattern stops looking for additional A or B events. The 'a' and 'b' tags retain only the first 3 (combined) matches among A and B events. The output event contains these tagged events as indexed properties.

5.5.4.5. Bound Repeat - Bounded Range

A bounded range specifies a low endpoint and a high endpoint.

The next pattern matches after at least one A event arrives upon the arrival of a single B event:

[1:3] a=A until B

If a B event arrives before the first A event, then the pattern does not match. Only the first 3 A events are returned by the pattern.

5.5.4.6. Tags and the Repeat Operator

The tags assigned to events in filter subexpressions within a repeat operator are available for use in filter expressions and also in any EPL clause.

This sample pattern matches 2 A events followed by a B event. Note the filter on B events: only a B event that has a value for the "beta" property that equals any of the "id" property values of the two A events is considered:

[2] A -> B(beta in (a[0].id, a[1].id))

The next EPL statement returns pairs of A events:

select a, a[0], a[0].id, a[1], a[1].id
from pattern [ every [2] a=A ] 

The select clause of the statement above showcases different ways of accessing tagged events:

  • The tag itself can be used to select an array of underlying events. For example, the 'a' expression above returns an array of underlying events of event type A.

  • The tag as an indexed property returns the underlying event at that index. For instance, the 'a[0]' expression returns the first underlying A event, or null if no such A event was matched by the repeat operator.

  • The tag as a nested, indexed property returns a property of the underlying event at that index. For example, the 'a[1].id' expression returns the 'id' property value of the second A event, or null if no such second A event was matched by the repeat operator.

You may not use indexed tags defined in the sub-expression to the repeat operator in the same subexpression. For example, in the following pattern the subexpression to the repeat operator is (a=A() -> b=B(id=a[0].id)) and the tag a cannot be used in its indexed form in the filter for event B:

// invalid
every [2] (a=A() -> b=B(id=a[0].id))

You can use tags without an index:

// valid
every [2] (a=A() -> b=B(id=a.id))

5.5.5. And

Similar to the C# && operator the and operator requires both nested pattern expressions to turn true before the whole expression turns true (a join pattern).

This pattern matches when both an A event and a B event arrive, at the time the last of the two events arrive:

A and B

This pattern matches on any sequence of an A event followed by a B event and then a C event followed by a D event, or a C event followed by a D and an A event followed by a B event:

(A -> B) and (C -> D)

Note that in an and pattern expression it is not possible to correlate events based on event property values. For example, this is an invalid pattern:

// This is NOT valid
a=A and B(id = a.id)

The above expression is invalid as it relies on the order of arrival of events, however in an and expression the order of events is not specified and events fulfill an and condition in any order. The above expression can be changed to use the followed-by operator:

// This is valid
a=A -> B(id = a.id)
// another example using 'and'...
a=A -> (B(id = a.id) and C(id = a.id))

Consider a pattern that looks for the same event:

A and A

The pattern above fires when a single A event arrives. The first arriving A event triggers a state transition in both the left and the right hand side expression.

In order to match after two A events arrive in any order, there are two options to express this pattern. The followed-by operator is one option and the repeat operator is the second option, as the next two patterns show:

A -> A
// ... or ...
[2] A

5.5.6. Or

Similar to the C# “||” operator the or operator requires either one of the expressions to turn true before the whole expression turns true.

Look for either an A event or a B event. As always, A and B can itself be nested expressions as well.

A or B

Detect all stock ticks that are either above or below a threshold.

every (StockTick(symbol='IBM', price < 100) or StockTick(symbol='IBM', price > 105)

5.5.7. Not

The not operator negates the truth value of an expression. Pattern expressions prefixed with not are automatically defaulted to true upon start, and turn permanently false when the expression within turns true.

The not operator is generally used in conjunction with the and operator or subexpressions as the below examples show.

This pattern matches only when an A event is encountered followed by a B event but only if no C event was encountered before either an A event and a B event, counting from the time the pattern is started:

(A -> B) and not C

Assume we'd like to detect when an A event is followed by a D event, without any B or C events between the A and D events:

A -> (D and not (B or C))

It may help your understanding to discuss a pattern that uses the or operator and the not operator together:

a=A -> (b=B or not C)

In the pattern above, when an A event arrives then the engine starts the subexpression B or not C. As soon as the subexpression starts, the not operator turns to true. The or expression turns true and thus your listener receives an invocation providing the A event in the property 'a'. The subexpression does not end and continues listening for B and C events. Upon arrival of a B event your listener receives a second invocation. If instead a C event arrives, the not turns permanently false however that does not affect the or operator (but would end an and operator).

To test for absence of an event, use timer:interval together with and not operators. The sample statement reports each 10-second interval during which no A event occurred:

every (timer:interval(10 sec) and not A)

In many cases the not operator, when used alone, does not make sense. The following example is invalid and will log a warning when the engine is started:

// not a sensible pattern
(not a=A) -> B(id=a.id)

5.5.8. Followed-by

The followed by -> operator specifies that first the left hand expression must turn true and only then is the right hand expression evaluated for matching events.

Look for an A event and if encountered, look for a B event. As always, A and B can itself be nested event pattern expressions.

A -> B

This is a pattern that fires when 2 status events indicating an error occur one after the other.

StatusEvent(status='ERROR') -> StatusEvent(status='ERROR')

A pattern that takes all A events that are not followed by a B event within 5 minutes:

every A -> (timer:interval(5 min) and not B)

A pattern that takes all A events that are not preceded by B within 5 minutes:

every (timer:interval(5 min) and not B -> A)

5.5.8.1. Limiting Sub-Expression Count

The followed-by -> operator can optionally be provided with an expression that limits the number of sub-expression instances of the right-hand side pattern sub-expression.

The synopsis for the followed-by operator with limiting expression is:

lhs_expression -[limit_expression]> rhs_expression

Each time the lhs_expression pattern sub-expression turns true the pattern engine starts a new rhs_expression pattern sub-expression. The limit_expression returns an integer value that defines a maximum number of pattern sub-expression instances that can simultaneously be present for the same rhs_expression.

When the limit is reached the pattern engine issues a notification to any condition handlers registered with the engine as described in Section 10.11, “Condition Handling” and does not start a new pattern sub-expression instance for the right-hand side pattern sub-expression.

For example, consider the following pattern which returns for every A event the first B event that matches the id field value of the A event:

every a=A -> b=B(id = a.id)

In the above pattern, every time an A event arrives (lhs) the pattern engine starts a new pattern sub-expression (rhs) consisting of a filter for the first B event that has the same value for the id field as the A event.

In some cases your application may want to limit the number of right-hand side sub-expressions because of memory concerns or to reduce output. You may add a limit expression returning an integer value as part of the operator.

This example employs the followed-by operator with a limit expression to indicate that maximally 2 filters for B events (the right-hand side pattern sub-expression) may be active at the same time:

every a=A -[2]> b=B(id = a.id)

Note that the limit expression in the example above is not a limit per value of id field, but a limit counting all right-hand side pattern sub-expression instances that are managed by that followed-by sub-expression instance.

If your followed-by operator lists multiple sub-expressions with limits, each limit applies to the immediate right-hand side. For example, the pattern below limits the number of filters for B events to 2 and the number of filters for C events to 3:

every a=A -[2]> b=B(id = a.id) -[3]> c=C(id = a.id)

5.5.9. Pattern Guards

Guards are where-conditions that control the lifecycle of subexpressions. Custom guard functions can also be used. The section Chapter 13, Extension and Plug-in outlines guard plug-in development in greater detail.

The pattern guard where-condition has no relationship to the EPL where clause that filters sets of events.

Take as an example the following pattern expression:

MyEvent where timer.within(10 sec)

In this pattern the timer:within guard controls the subexpression that is looking for MyEvent events. The guard terminates the subexpression looking for MyEvent events after 10 seconds after start of the pattern. Thus the pattern alerts only once when the first MyEvent event arrives within 10 seconds after start of the pattern.

The every keyword requires additional discussion since it also controls subexpression lifecycle. Let's add the every keyword to the example pattern:

every MyEvent where timer.within(10 sec)

The difference to the pattern without every is that each MyEvent event that arrives now starts a new subexpression, including a new guard, looking for a further MyEvent event. The result is that, when a MyEvent arrives within 10 seconds after pattern start, the pattern execution will look for the next MyEvent event to arrive within 10 seconds after the previous one.

By placing parentheses around the every keyword and its subexpression, we can have the every under the control of the guard:

(every MyEvent) where timer.within(10 sec)

In the pattern above, the guard terminates the subexpression looking for all MyEvent events after 10 seconds after start of the pattern. This pattern alerts for all MyEvent events arriving within 10 seconds after pattern start, and then stops.

Guards do not change the truth value of the subexpression of which the guard controls the lifecycle, and therefore do not cause a restart of the subexpression when used with the every operator. For example, the next pattern stops returning matches after 10 seconds unless a match occurred within 10 seconds after pattern start:

every ( (A and B) where timer.within(10 sec) )

5.5.9.1. The timer:within Pattern Guard

The timer:within guard acts like a stopwatch. If the associated pattern expression does not turn true within the specified time period it is stopped and permanently false.

The synopsis for timer:within is as follows:

timer:within(time_period_expression)

The time_period_expression is a time period (see Section 4.2.1, “Specifying Time Periods”) or an expression providing a number of seconds as a parameter. The interval expression may contain references to properties of prior events in the same pattern as well as variables and substitution parameters.

This pattern fires if an A event arrives within 5 seconds after statement creation.

A where timer:within (5 seconds)

This pattern fires for all A events that arrive within 5 seconds. After 5 seconds, this pattern stops matching even if more A events arrive.

(every A) where timer:within (5 seconds)

This pattern matches for any one A or B event in the next 5 seconds.

( A or B ) where timer:within (5 sec)

This pattern matches for any 2 errors that happen 10 seconds within each other.

every (StatusEvent(status='ERROR') -> StatusEvent(status='ERROR') where timer:within (10 sec))

The following guards are equivalent:

timer:within(2 minutes 5 seconds)
timer:within(125 sec)
timer:within(125)

5.5.9.2. The timer:withinmax Pattern Guard

The timer:withinmax guard is similar to the timer:within guard and acts as a stopwatch that additionally has a counter that counts the number of matches. It ends the subexpression when either the stopwatch ends or the match counter maximum value is reached.

The synopsis for timer:withinmax is as follows:

timer:withinmax(time_period_expression, max_count_expression)

The time_period_expression is a time period (see Section 4.2.1, “Specifying Time Periods”) or an expression providing a number of seconds.

The max_count_expression provides the maximum number of matches before the guard ends the subexpression.

Each parameter expression may also contain references to properties of prior events in the same pattern as well as variables and substitution parameters.

This pattern fires for every A event that arrives within 5 seconds after statement creation but only up to the first two A events:

(every A) where timer:withinmax (5 seconds, 2)

If the result of the max_count_expression is 1, the guard ends the subexpression after the first match and indicates the first match.

This pattern fires for the first A event that arrives within 5 seconds after statement creation:

(every A) where timer:withinmax (5 seconds, 1)

If the result of the max_count_expression is zero, the guard ends the subexpression upon the first match and does no indicate any matches.

This example receives every A event followed by every B event (as each B event arrives) until the 5-second subexpression timer ends or X number of B events have arrived (assume X was declared as a variable):

every A -> (every B) where timer:withinmax (5 seconds, X)

5.5.9.3. The while Pattern Guard

The while guard is followed by an expression that the engine evaluates for every match reported by the guard pattern sub-expression. When the expression returns false the pattern sub-expression ends.

The synopsis for while is as follows:

while (guard_expression)

The guard_expression is any expression that returns a boolean true or false. The expression may contain references to properties of prior events in the same pattern as well as variables and substitution parameters.

Each time the subexpression indicates a match, the engine evaluates guard_expression and if true, passes the match and when false, ends the subexpression.

This pattern fires for every A event until an A event arrives that has a value of zero or less for its size property (assuming A events have an integer size property).

(every a=A) while (a.size > 0)

Note the parenthesis around the every subexpression. They ensure that, following precedence rules, the guard applies to the every operator as well.

5.5.9.4. Guard Time Interval Expressions

The timer:within and timer:withinmax guards may be parameterized by an expression that contains one or more references to properties of prior events in the same pattern.

As a simple example, this pattern matches every A event followed by a B event that arrives within delta seconds after the A event:

every a=A -> b=B where timer:within (a.delta seconds)

Herein A event is assumed to have a delta property that provides the number of seconds to wait for B events. Each arriving A event may have a different value for delta and the guard is therefore parameterized dynamically based on the prior A event received.

When multiple events accumulate, for example when using the match-until or repeat pattern elements, an index must be provided:

[2] a=A -> b=B where timer:within (a[0].delta + a[1].delta)

The above pattern matches after 2 A events arrive followed by a B event within a time interval after the A event that is defined by the sum of the delta properties of both A events.

5.5.9.5. Combining Guard Expressions

You can combine guard expression by using parenthesis around each subexpression.

The below pattern matches for each A event while A events of size greater then zero arrive and only within the first 20 seconds:

((every a=A) while (a.size > 0)) where timer:within(20)

5.6. Pattern Atoms

5.6.1. Filter Atoms

Filter atoms have been described in section Section 5.4, “Filter Expressions In Patterns”.

5.6.2. Time-based Observer Atoms

Observers observe time-based events for which the thread-of-control originates by the engine timer or external timer event. Custom observers can also be developed that observe timer events or other engine-external application events such as a file-exists check. The section Chapter 13, Extension and Plug-in outlines observer plug-in development in greater detail.

5.6.2.1. timer:interval

The timer:interval observer waits for the defined time before the truth value of the observer turns true. The observer takes a time period (see Section 4.2.1, “Specifying Time Periods”) as a parameter, or an expression that returns the number of seconds.

The observer may be parameterized by an expression that contains one or more references to properties of prior events in the same pattern, or may also reference variables, substitution parameters or any other expression returning a numeric value.

After an A event arrived wait 10 seconds then indicate that the pattern matches.

A -> timer:interval(10 seconds) 

The pattern below fires every 20 seconds.

every timer:interval(20 sec)

The next example pattern fires for every A event that is not followed by a B event within 60 seconds after the A event arrived. The B event must have the same "id" property value as the A event.

every a=A -> (timer:interval(60 sec) and not B(id=a.id)) 

Consider the next example, which assumes that the A event has a property waittime:

every a=A -> (timer:interval(a.waittime + 2) and not B(id=a.id))

In the above pattern the logic waits for 2 seconds plus the number of seconds provided by the value of the waittime property of the A event.

5.6.2.2. timer:at

The timer:at observer is similar in function to the Unix “crontab” command. At a specified time the expression turns true. The at operator can also be made to pattern match at regular intervals by using an every operator in front of the timer:at operator.

The syntax is: timer:at (minutes, hours, days of month, months, days of week [, seconds]).

The value for seconds is optional. Each element allows wildcard * values. Ranges can be specified by means of lower bounds then a colon ‘:’ then the upper bound. The division operator */x can be used to specify that every xth value is valid. Combinations of these operators can be used by placing these into square brackets([]).

The timer:at observer may also be parameterized by an expression that contains one or more references to properties of prior events in the same pattern, or may also reference variables, substitution parameters or any other expression returning a numeric value. The frequency division operator */x and parameters lists within brackets([]) are an exception: they may only contain variables, substitution parameters or numeric values.

This expression pattern matches every 5 minutes past the hour.

every timer:at(5, *, *, *, *)

The below timer:at pattern matches every 15 minutes from 8am to 5:45pm (hours 8 to 17 at 0, 15, 30 and 45 minutes past the hour) on even numbered days of the month as well as on the first day of the month.

timer:at (*/15, 8:17, [*/2, 1], *, *)

The below table outlines the fields, valid values and keywords available for each field:

Table 5.5. Properties offered by sample statement aggregating price

Field NameMandatory?Allowed ValuesAdditional Keywords
Minutesyes0 - 59 
Hoursyes0 - 23 
Days Of Monthyes1 - 31last, weekday, lastweekday
Monthsyes1 - 12 
Days Of Weekyes0 (Sunday) - 6 (Saturday)last
Secondsno0 - 59 

The keyword last used in the days-of-month field means the last day of the month (current month). To specify the last day of another month, a value for the month field has to be provided. For example: timer:at(*, *, last,2,*) is the last day of February.

The last keyword in the day-of-week field by itself simply means Saturday. If used in the day-of-week field after another value, it means "the last xxx day of the month" - for example "5 last" means "the last friday of the month". So the last Friday of the current month will be: timer:at(*, *, *, *, 5 last). And the last Friday of June: timer:at(*, *, *, 6, 5 last).

The keyword weekday is used to specify the weekday (Monday-Friday) nearest the given day. Variant could include month like in: timer:at(*, *, 30 weekday, 9, *) which is Friday September 28th (no jump over month).

The keyword lastweekday is a combination of two parameters, the last and the weekday keywords. A typical example could be: timer:at(*, *, *, lastweekday, 9, *) which will define Friday September 28th (example year is 2007).

5.6.2.2.1. timer:at and the every Operator

When using timer:at with the every operator the crontab-like timer computes the next time at which the timer should fire based on the specification and the current time. When using every, the current time is the time the timer fired or the statement start time if the timer has not fired once.

For example, this pattern fires every 1 minute starting at 1:00pm and ending at 1:59pm, every day:

every timer:at(*, 13, *, *, *)

Assume the above statement gets started at 1:05pm and 20 seconds. In such case the above pattern fires every 1 minute starting at 1:06pm and ending at 1:59pm for that day and 1:00pm to 1:59pm every following day.

To get the pattern to fire only once at 1pm every day, explicitly specify the minute to start. The pattern below fires every day at 1:00pm:

every timer:at(0, 13, *, *, *)

By specifying a second resolution the timer can be made to fire every second, for instance:

every timer:at(*, *, *, *, *, *)

Chapter 6. EPL Reference: Match Recognize

6.1. Overview

Using match recognize patterns are defined in the familiar syntax of regular expressions.

The match recognize syntax presents an alternative way to specify pattern detection as compared to the EPL pattern language described in the previous chapter. A comparison of match recognize and EPL patterns is below.

The match recognize syntax is a proposal for incorporation into the SQL standard. It is thus subject to change as the standard evolves and finalizes (it has not finalized yet). Please consult row-pattern-recogniton-11-public for further information.

You may be familiar with regular expressions in the context of finding text of interest in a string, such as particular characters, words, or patterns of characters. Instead of matching characters, match recognize matches sequences of events of interest.

Esper can apply match-recognize patterns in real-time upon arrival of new events in a stream of events (also termed incrementally, streaming or continuous). Esper can also match patterns on-demand via the iterator pull-API, if specifying a named window or data window on a stream.

6.2. Comparison of Match Recognize and EPL Patterns

This section compares pattern detection via match recognize and via the EPL pattern language.

Table 6.1. Comparison Match Recognize to EPL Patterns

CategoryEPL PatternsMatch Recognize
PurposePattern detection in sequences of events.Same.
StandardsNot standardized, similar to Rapide pattern language.Proposal for incorporation into the SQL standard.
Real-time ProcessingYes.Yes.
On-Demand query via IteratorNo.Yes.
LanguageNestable expressions consisting of boolean AND, OR, NOT and time or arrival-based constructs such as -> (followed-by), timer:within and timer:interval.Regular expression consisting of variables each representing conditions on events.
Event TypesAn EPL pattern may react to multiple different types of events.The input is a single type of event (unless used with variant streams).
Data Window InteractionDisconnected, i.e. an event leaving a data window does not change pattern state.Connected, i.e. an event leaving a data window removes the event from match selection.
Semantic EvaluationTruth-value based: A EPL pattern such as (A and B) can fire when a single event arrives that satisfies both A and B conditions.Sequence-based: A regular expression (A B) requires at least two events to match.
Time Relationship Between EventsThe timer:within, timer:interval and NOT operator can expressively search for absence of events or other more complex timing relationships.Some support for detecting absence of events using the interval clause.
ExtensibilityCustom pattern objects, user-defined functions.User-defined functions, custom aggregation functions.
Memory UseLikely between 500 bytes to 2k per open sequence, depends on pattern.Likely between 100 bytes to 1k per open sequence, depends on pattern.

6.3. Syntax

The synopsis is as follows:

match_recognize (
  [ partition by partition_expression [, partition_expression] [,...]  ]
  measures  measure_expression as col_name [, measure_expression as col_name ] [,...]
  [ all matches ]
  [ after match skip (past last row | to next row | to current row) ]
  pattern ( variable_regular_expr [, variable_regular_expr] [,...] )
  [ interval time_period ]
  define  variable as variable_condition [, variable as variable_condition]  [,...] 
)

The match_recognize keyword starts the match recognize definition and occurs right after the from clause in an EPL select statement. It is followed by parenthesis that surround the match recognize definition.

Partition by is optional and may be used to specify that events are to be partitioned by one or more event properties or expressions. If there is no Partition by then all rows of the table constitute a single partition. The regular expression applies to events in the same partition and not across partitions.

The measures clause defines columns that contain expressions over the pattern variables. The expressions can reference partition columns, singleton variables, aggregates as well as indexed properties on the group variables. Each measure_expression expression must be followed by the as keyword and a col_name column name.

The all matches keywords are optional and instructs the engine to find all possible matches. By default matches are ranked and the engine returns a single match following an algorithm to eliminate duplicate matches, as described below. When specifying all matches, matches may overlap and may start at the same row.

The after match skip keywords are optional and serve to determine the resumption point of pattern matching after a match has been found. By default the behavior is after match skip past last row. This means that after eliminating duplicate matches, the logic skips to resume pattern matching at the next event after the last event of the current match.

The pattern component is used to specify a regular expression. The regular expression is built from variable names, and may use the operators such as *, +, ?, *?, +?, ?? quantifiers and | alteration (concatenation is indicated by the absence of any operator sign between two successive items in a pattern).

With the optional interval keyword and time period you can control how long the engine should wait for further events to arrive that may be part of a matching event sequence, before indicating a match (or matches) (not applicable to on-demand pattern matching).

Define is a mandatory component, used to specify the boolean condition that defines a variable name that is declared in the pattern. A variable name does not require a definition and if there is no definition, the default is a predicate that is always true. Such a variable name can be used to match any row.

6.3.1. Syntax Example

For illustration, the examples in this chapter use the TemperatureSensorEvent event. Each event has 3 properties: the id property is a unique event id, the device is a sensor device number and the temp property is a temperature reading. An event described as "id=E1, device=1, temp=100" is a TemperatureSensorEvent event with id "E1" for device 1 with a reading of 100.

This example statement looks for two TemperatureSensorEvent events from the same device, directly following each other, that indicate a jump in temperature of 10 or more between the two events:

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures A.id as a_id, B.id as b_id, A.temp as a_temp, B.temp as b_temp
  pattern (A B)
  define 
    B as Math.abs(B.temp - A.temp) >= 10
)

The partition by ensures that the regular expression applies to sequences of events from the same device.

The measures clause provides a list of properties or expressions to be selected from matching events. Each property name must be prefixed by the variable name.

In the pattern component the statement declares two variables: A and B. As a matter of convention, variable names are uppercase characters.

The define clause specifies no condition for variable A. This means that A defaults to true and any event matches A. Therefore, the pattern can start at any event.

The pattern A B indicates to look for a pattern in which an event that fulfills the condition for variable A is immediately followed by an event that fulfills the condition for variable B. A pattern thus defines the sequence (or sequences) of conditions that must be met for the pattern to fire.

Below table is an example sequence of events and output of the pattern:

Table 6.2. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=50 
2000id=E2, device=1, temp=55 
3000id=E3, device=1, temp=60 
4000id=E4, device=1, temp=70a_id = E3, b_id = E4, a_temp = 60, b_temp = 70
5000id=E5, device=1, temp=85 
6000id=E6, device=1, temp=85 
7000id=E7, device=2, temp=100 

At time 4000 when event with id E4 (or event E4 or just E4 for short) arrives the pattern matches and produces an output event. Matching then skips past the last event of the current match (E4) and begins at event E5 (the default skip clause is past last row). Therefore events E4 and E5 do not constitute a match.

At time 3000, events E1 and E3 do not constitute a match as E3 does not immediately follow E, since there is E2 in between.

At time 7000, event E7 does not constitute a match as it is from device 2 and thereby not in the same partition as prior events.

6.4. Pattern and Pattern Operators

The pattern specifies the pattern to be recognized in the ordered sequence of events in a partition using regular expression syntax. Each variable name in a pattern corresponds to a boolean condition, which is specified later using the define component of the syntax. Thus the pattern can be regarded as implicitly declaring one or more variable names; the definition of those variable names appears later in the syntax. If a variable is not defined the variable defaults to true.

It is permitted for a variable name to occur more than once in a pattern, for example pattern (A B A).

6.4.1. Operator Precedence

The operators at the top of this table take precedence over operators lower on the table.

Table 6.3. Match Recognize Regular Expression Operator Precedence

PrecedenceOperatorDescriptionExample
1Grouping()
(A B)
2Single-character duplication* + ?
A* B+ C?
3Concatenation(no operator)
A B
4Alternation|
A | B

If you are not sure about the precedence, please consider placing parenthesis () around your groups. Parenthesis can also help make expressions easier to read and understand.

6.4.2. Concatenation

The concatenation is indicated by the absence of any operator sign between two successive items in a pattern.

In below pattern the two items A and B have no operator between them. The pattern matches for any event immediately followed by an event from the same device that indicates a jump in temperature over 10:

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures A.id as a_id, B.id as b_id, A.temp as a_temp, B.temp as b_temp
  pattern (A B)
  define 
    B as Math.abs(B.temp - A.temp) >= 10
)

Please see the Section 6.3.1, “Syntax Example” for a sample event sequence.

6.4.3. Alternation

The alternation operator is a vertical bar ( | ).

The alternation operator has the lowest precedence of all operators. It tells the engine to match either everything to the left of the vertical bar, or everything to the right of the vertical bar. If you want to limit the reach of the alternation, you will need to use round brackets for grouping.

This example pattern looks for a sequence of an event with a temperature over 50 followed immediately by either an event with a temperature less then 45 or an event that indicates the temperature jumped by 10 (all for the same device):

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures A.id as a_id, B.id as b_id, C.id as c.id
  pattern (A (B | C))
  define 
    A as A.temp >= 50,
    B as B.temp <= 45,
    C as Math.abs(B.temp - A.temp) >= 10)

Below table is an example sequence of events and output of the pattern:

Table 6.4. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=50 
2000id=E2, device=1, temp=45a_id=E1, b_id=E2, c_id=null
3000id=E3, device=1, temp=46 
4000id=E4, device=1, temp=48 
5000id=E5, device=1, temp=50 
6000id=E6, device=1, temp=60a_id = E5, b_id = null, c_id=E6

6.4.4. Quantifiers Overview

Quantifiers are postfix operators with the following choices:

Table 6.5. Quantifiers

QuantifierMeaning
*Zero or more matches (greedy).
+One or more matches (greedy).
?Zero or one match (greedy).
*?Zero or more matches (reluctant).
+?One or more matches (reluctant).
??Zero or one match (reluctant).

6.4.5. Variables Can be Singleton or Group

A singleton variable is a variable in a pattern that does not have a quantifier or has a zero-or-one quantifier (? or ??) and occurs only once in the pattern (except with alteration). In the measures clause a singleton variable can be selected as:

variableName.propertyName

Variables with a zero-or-more or one-or-more quantifier, or variables that occur multiple places in a pattern (except when using alteration), may match multiple events and are group variables. In the measures clause a group variable must be selected either by providing an index or via any of the aggregation functions, such as first, last, count and sum:

variableName[index].propertyName
last(variableName.propertyName)

Please find examples of singleton and group variables and example measures clauses below.

6.4.5.1. Additional Aggregation Functions

For group variables all existing aggregation functions can be used and in addition the following aggregation functions may be used:

Table 6.6. Syntax and results of aggregate functions

Aggregate FunctionResult
first([all|distinct] expression)

Returns the first value.

last([all|distinct] expression)

Returns the last value.

6.4.6. Eliminating Duplicate Matches

The execution of match recognize is continuous and real-time by default. This means that every arriving event, or batch of events if using batching, evaluates against the pattern and matches are immediately indicated. Elimination of duplicate matches occurs between all matches of the arriving events (or batch of events) at a given time.

As an alternative, and if your application does not require continuous pattern evaluation, you may use the iterator API to perform on-demand matching of the pattern. For the purpose of indicating to the engine to not generate continuous results, specify the @Hint('iterate_only') hint.

When using one-or-more, zero-or-more or zero-or-one quantifiers (?, +, *, ??, +?, *?), the output of the real-time continuous query can differ from the output of the on-demand iterator execution: The continuous query will output a match (or multiple matches) as soon as matches are detected at a given time upon arrival of events (not knowing what further events may arrive). The on-demand execution, since it knows all possible events in advance, can determine the longest match(es). Thus elimination of duplicate matches can lead to different results between real-time and on-demand use.

If the all matches keywords are specified, then all matches are returned as the result and no elimination of duplicate matches as below occurs.

Otherwise matches to a pattern in a partition are ordered by preferment. Preferment is given to matches based on the following priorities:

  1. A match that begins at an earlier row is preferred over a match that begins at a later row.

  2. Of two matches matching a greedy quantifier, the longer match is preferred.

  3. Of two matches matching a reluctant quantifier, the shorter match is preferred.

After ranking matches by preferment, matches are chosen as follows:

  1. The first match by preferment is taken.

  2. The pool of matches is reduced as follows based on the SKIP TO clause: If SKIP PAST LAST ROW is specified, all matches that overlap the first match are discarded from the pool. If SKIP TO NEXT ROW is specified, then all matches that overlap the first row of the first match are discarded. If SKIP TO CURRENT ROW is specified, then no matches are discarded.

  3. The first match by preferment of the ones remaining is taken.

  4. Step 2 is repeated to remove more matches from the pool.

  5. Steps 3 and 4 are repeated until there are no remaining matches in the pool.

6.4.7. Greedy Or Reluctant

Reluctant quantifiers are indicated by an additional question mark (*?, +?, ??,). Reluctant quantifiers try to match as few rows as possible, whereas non-reluctant quantifiers are greedy and try to match as many rows as possible.

Greedy and reluctant come into play only for match selection among multiple possible matches. When specifying all matches there is no difference between greedy and reluctant quantifiers.

Consider the below example. The conditions may overlap: an event with a temperature reading of 105 and over matches both A and B conditions:

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures A.id as a_id, B.id as b_id
  pattern (A?? B?)
  define 
    A as A.temp >= 100
    B as B.temp >= 105)

A sample sequence of events and pattern matches:

Table 6.7. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=99 
2000id=E2, device=2, temp=106a_id=null, b_id=E2
3000id=E3, device=1, temp=100a_id=E3, b_id=null

As the ? qualifier on condition B is greedy, event E2 matches the pattern and is indicated as a B event by the measure clause (and not as an A event therefore a_id is null).

6.4.8. Quantifier - One Or More (+ and +?)

The one-or-more quantifier (+) must be matched one or more times by events. The operator is greedy and the reluctant version is +?.

In the below example with pattern (A+ B+) the pattern consists of two variable names, A and B, each of which is quantified with +, indicating that they must be matched one or more times.

The pattern looks for one or more events in which the temperature is over 100 followed by one or more events indicating a higher temperature:

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures first(A.id) as first_a, last(A.id) as last_a, B[0].id as b0_id, B[1].id as b1_id
  pattern (A+ B+)
  define 
	A as A.temp >= 100,
	B as B.temp > A.temp)

An example sequence of events that matches the pattern above is:

Table 6.8. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=99 
2000id=E2, device=1, temp=100 
3000id=E3, device=1, temp=100 
4000id=E4, device=1, temp=101first_a = E2, last_a = E3, b0_id = E4, b1_id = null
5000id=E5, device=1, temp=102 

Note that for continuous queries, there is no match that includes event E5 since after the pattern matches for E4 the pattern skips to start fresh at E5 (by default skip clause). When performing on-demand matching via iterator, event E5 gets included in the match and the output is first_a = E2, last_a = E3, b0_id = E4, b1_id = E5.

6.4.9. Quantifier - Zero Or More (* and *?)

The zero-or-more quantifier (*) must be matched zero or more times by events. The operator is greedy and the reluctant version is *?.

The pattern looks for a sequence of events in which the temperature starts out below 50 and then stays between 50 and 60 and finally comes over 60:

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures A.id as a_id, count(B.id) as count_b, C.id as c_id
  pattern (A B* C)
  define 
	A as A.temp < 50,
	B as B.temp between 50 and 60,
	C as C.temp > 60)

An example sequence of events that matches the pattern above is:

Table 6.9. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=55 
2000id=E2, device=1, temp=52 
3000id=E3, device=1, temp=49 
4000id=E4, device=1, temp=51 
5000id=E5, device=1, temp=55 
6000id=E5, device=1, temp=61a_id=E3, count_b=2, c_id=E6

6.4.10. Quantifier - Zero Or One (? and ??)

The zero-or-one quantifier (?) must be matched zero or one time by events. The operator is greedy and the reluctant version is ??.

The pattern looks for a sequence of events in which the temperature is below 50 and then dips to over 50 and then to under 50 before indicating a value over 55:

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures A.id as a_id, B.id as b_id, C.id as c_id, D.id as d_id
  pattern (A B? C? D)
  define 
	A as A.temp < 50,
	B as B.temp > 50,
	C as C.temp < 50,
	D as D.temp > 55)

An example sequence of events that matches the pattern above is:

Table 6.10. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=44 
2000id=E2, device=1, temp=49 
3000id=E3, device=1, temp=51 
4000id=E4, device=1, temp=49 
5000id=E5, device=1, temp=56a_id=E2, b_id=E3, c_id=E4, d_id=E5
6000id=E5, device=1, temp=61 

6.5. Define Clause

Within define are listed the boolean conditions that defines a variable name that is declared in the pattern.

A variable name does not require a definition and if there is no definition, the default is a predicate that is always true. Such a variable name can be used to match any row.

The definitions of variable names may reference the same or other variable names as prior examples have shown.

If a variable in your condition expression is a singleton variable, then only individual columns may be referenced. If the variable is not matched by an event, a null value is returned.

If a variable in your condition expression is a group variable, then only indexed columns may be referenced. If the variable is not matched by an event, a null value is returned.

Aggregation functions are not allowed within expressions of the define clause.

6.5.1. The Prev Operator

The prev function may be used in a define expression to access columns of the previous row of a variable name. If there is no previous row, the null value is returned.

The prev function can accept an optional non-negative integer argument indicating the offset to the previous rows. That argument must be a constant. In this case, the engine returns the property from the N-th row preceding the current row, and if the row doesn’t exist, it returns null.

This function can access variables currently defined, for example:

Y as Y.price < prev(Y.price, 2)

It is not legal to use prev with another variable then the one being defined:

// not allowed
Y as Y.price < prev(X.price, 2)

The prev function returns properties of events in the same partition. Also, it returns properties of events according to event order-of-arrival. When using data windows or deleting events from a named window, the remove stream does not remove events from the prev function.

The pattern looks for an event in which the temperature is greater or equal 100 and that, relative to that event, has an event preceding it by 2 events that also had a temperature greater or equal 100:

select * from TemperatureSensorEvent
match_recognize (
  partition by device
  measures A.id as a_id
  pattern (A)
  define 
	A as A.temp > 100 and prev(A.temp, 2) > 100)

An example sequence of events that matches the pattern above is:

Table 6.11. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=98 
2000id=E2, device=1, temp=101 
3000id=E3, device=1, temp=101 
4000id=E4, device=1, temp=99 
5000id=E5, device=1, temp=101a_id=E5

6.6. Measure Clause

The measures clause defines exported columns that contain expressions over the pattern variables. The expressions can reference partition columns, singleton variables and any aggregation functions including last and first on the group variables.

Expressions in the measures clause must use the as keyword to assign a column name.

If a variable is a singleton variable then only individual columns may be referenced, not aggregates. If the variable is not matched by an event, a null value is returned.

If a variable is a group variable and used in an aggregate, then the aggregate is performed over all rows that have matched the variable. If a group variable is not in an aggregate function, its variable name must be post-fixed with an index. See Section 6.4.5, “Variables Can be Singleton or Group” for more information.

6.7. Datawindow-Bound

When using match recognize with a named window or stream bound by a data window, all events removed from the named window or data window also removed the match-in-progress that includes the event(s) removed.

The next example looks for four sensor events from the same device immediately following each other and indicating a rising temperature, but only events that arrived in the last 10 seconds are considered:

select * from TemperatureSensorEvent.win:time(10 sec)
match_recognize (
partition by device
measures A.id as a_id
pattern (A B C D)
define 
B as B.temp > A.temp,
C as C.temp > B.temp,
D as D.temp > C.temp)

An example sequence of events that matches the pattern above is:

Table 6.12. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=80 
2000id=E2, device=1, temp=81 
3000id=E3, device=1, temp=82 
4000id=E4, device=1, temp=81 
7000id=E5, device=1, temp=82 
9000id=E6, device=1, temp=83 
13000id=E7, device=1, temp=84a_id=E4, a_id=E5, a_id=E6, a_id=E7
15000id=E8, device=1, temp=84 
20000id=E9, device=1, temp=85 
21000id=E10, device=1, temp=86 
26000id=E11, device=1, temp=87 

Note that E8, E9, E10 and E11 doe not constitute a match since E8 leaves the data window at 25000.

6.8. Interval

With the optional interval keyword and time period you can control how long the engine should wait for further events to arrive that may be part of a matching event sequence, before indicating a match (or matches). This is not applicable to on-demand pattern matching.

The interval timer starts are the arrival of the first event matching a sequence for a partition. When the time interval passes and an event sequence matches, duplicate matches are eliminated and output occurs.

The next example looks for sensor events indicating a temperature of over 100 waiting for any number of additional events with a temperature of over 100 for 10 seconds before indicating a match:

select * from TemperatureSensorEvent
match_recognize (
partition by device
measures A.id as a_id, count(B.id) as count_b, first(B.id) as first_b, last(B.id) as last_b
pattern (A B*)
interval 5 seconds
define 
  A as A.temp > 100,
  B as B.temp > 100)

An example sequence of events that matches the pattern above is:

Table 6.13. Example

Arrival TimeTupleOutput Event (if any)
1000id=E1, device=1, temp=98 
2000id=E2, device=1, temp=101 
3000id=E3, device=1, temp=102 
4000id=E4, device=1, temp=104 
5000id=E5, device=1, temp=104 
7000 a_id=E2, count_b=3, first_b=E3, last_b=E5

Notice that the engine waits 5 seconds (5000 milliseconds) after the arrival time of the first event E2 of the match at 2000, to indicate the match at 7000.

6.9. Limitations

Please note the following limitations:

  1. Subqueries are not allowed in expressions within match_recognize.

  2. Joins and outer joins are not allowed in the same statement as match_recognize.

  3. match_recognize may not be used within on-select or on-insert statements.

  4. When using match_recognize on unbound streams (no data window provided) the iterator pull API returns no rows.

  5. A Statement Object Model API for match_recognize is not yet available.

Chapter 7. EPL Reference: Operators

Esper arithmetic and logical operator precedence follows standard arithmetic and logical operator precedence.

7.1. Arithmetic Operators

The below table outlines the arithmetic operators available.

Table 7.1. Syntax and results of arithmetic operators

OperatorDescription
+, -

As unary operators they denote a positive or negative expression. As binary operators they add or subtract.

*, /

Multiplication and division are binary operators.

%

Modulo binary operator.

7.2. Logical And Comparison Operators

The below table outlines the logical and comparison operators available.

Table 7.2. Syntax and results of logical and comparison operators

OperatorDescription
NOT

Returns true if the following condition is false, returns false if it is true.

OR

Returns true if either component condition is true, returns false if both are false.

AND

Returns true if both component conditions are true, returns false if either is false.

=, !=, <, > <=, >=,

Comparison.

7.3. Concatenation Operators

The below table outlines the concatenation operators available.

Table 7.3. Syntax and results of concatenation operators

OperatorDescription
||

Concatenates character strings

7.4. Binary Operators

The below table outlines the binary operators available.

Table 7.4. Syntax and results of binary operators

OperatorDescription
&

Bitwise AND if both operands are numbers; conditional AND if both operands are boolean.

|

Bitwise OR if both operands are numbers; conditional OR if both operands are boolean.

^

Bitwise exclusive OR (XOR).

7.5. Array Definition Operator

The { and } curly braces are array definition operators following the array initialization syntax. Arrays can be useful to pass to user-defined functions or to select array data in a select clause.

Array definitions consist of zero or more expressions within curly braces. Any type of expression is allowed within array definitions including constants, arithmetic expressions or event properties. This is the syntax of an array definition:

{ [expression [,expression...]] }

Consider the next statement that returns an event property named actions. The engine populates the actions property as an array of System.String values with a length of 2 elements. The first element of the array contains the observation property value and the second element the command property value of RFIDEvent events.

select {observation, command} as actions from RFIDEvent

The engine determines the array type based on the types returned by the expressions in the array definiton. For example, if all expressions in the array definition return integer values then the type of the array is System.Int32[]. If the types returned by all expressions are compatible number types, such as integer and double values, the engine coerces the array element values and returns a suitable type, System.Double[] in this example. The type of the array returned is Object[] if the types of expressions cannot be coerced or return object values. Null values can also be used in an array definition.

Arrays can come in handy for use as parameters to user-defined functions:

select * from RFIDEvent where Filter.myFilter(zone, {1,2,3})

7.6. Dot Operator

You can use the dot operator to invoke a method on the result of an expression. The dot operator uses the dot (.) or period character.

The synopsis for the dot operator is as follows

(expression).method([parameter [,...]])[.method(...)][...]

The expression to evaluate by the dot operator is in parenthesis. After the dot character follows the method name and method parameters in parenthesis.

You may use the dot operator when your expression returns an object that you want to invoke a method on. The dot operator allows duck typing and convenient array and collection access methods.

This example statement invokes the GetZones method of the RFID event class by referring to the stream name assigned in the from-clause:

select (rfid).GetZones() from RFIDEvent as rfid

The size() method can be used to return the array length or collection size. Use the Get method to return the value at a given index for an array or collection.

The next statement selects array size and returns the last array element:

select (arrayproperty).size() as arraySize, 
  (arrayproperty).get((arrayproperty).size - 1) as lastInArray 
  from ProductEvent

7.6.1. Duck Typing

Duck typing is when the engine checks at runtime for the existence of a method regardless of object class inheritance hierarchies. This can be useful, for example, when a dynamic property returns an object which may or may not provide a method to return the desired value.

Duck typing is disabled in the default configuration to consistently enforce strong typing. Please enable duck typing via engine expression settings as described in Section 11.4.18, “Engine Settings related to Expression Evaluation”.

The statement below selects a dynamic property by name productDesc and invokes the getCounter() method if that method exists on the property value, or returns the null value if the method does not exist for the dynamic property value of if the dynamic property value itself is null:

select (productDesc?).getCounter() as arraySize from ProductEvent

7.7. The 'in' Keyword

The in keyword determines if a given value matches any value in a list. The syntax of the keyword is:

test_expression [not] in (expression [,expression...] )

The test_expression is any valid expression. The keyword is followed by a list of expressions to test for a match. The optional not keyword specifies that the result of the predicate be negated.

The result of an in expression is of type Boolean. If the value of test_expression is equal to any expression from the comma-separated list, the result value is true. Otherwise, the result value is false.

The next example shows how the in keyword can be applied to select certain command types of RFID events:

select * from RFIDEvent where command in ('OBSERVATION', 'SIGNAL')

The statement is equivalent to:

select * from RFIDEvent where command = 'OBSERVATION' or command = 'SIGNAL'

Expression may also return an array, a System.Collection.Generic.ICollection or a System.Collection.Generic.IDictionary. Thus event properties that are lists, sets or maps may provide values to compare against test_expression.

All expressions must be of the same type or a compatible type to test_expression. The in keyword may coerce number values to compatible types. If expression returns an array, then the component type of the array must be compatible, unless the component type of the array is Object.

If expression returns an array of component type Object, the operation compares each element of the array, applying equals semantics.

If expression returns a Collection, the operation determines if the collection contains the value returned by test_expression, applying contains semantics.

If expression returns a Map, the operation determines if the map contains the key value returned by test_expression, applying containsKey semantics.

Constants, arrays, Collection and Map expressions or event properties can be used combined.

For example, and assuming a property named 'mySpecialCmdList' exists that contains a list of command strings:

select * from RFIDEvent where command in ( 'OBSERVATION', 'SIGNAL', mySpecialCmdList)

When using prepared statements and substitution parameters with the in keyword, make sure to retain the parenthesis. Substitution values may also be arrays, Collection and Map values:

test_expression [not] in (? [,?...] )

Note that if there are no successes and at least one right-hand row yields null for the operator's result, the result of the any construct will be null, not false. This is in accordance with SQL's normal rules for Boolean combinations of null values.

7.8. The 'between' Keyword

The between keyword specifies a range to test. The syntax of the keyword is:

test_expression [not] between begin_expression and end_expression

The test_expression is any valid expression and is the expression to test for in the range defined by begin_expression and end_expression. The not keyword specifies that the result of the predicate be negated.

The result of a between expression is of type Boolean. If the value of test_expression is greater then or equal to the value of begin_expression and less than or equal to the value of end_expression, the result is true.

The next example shows how the between keyword can be used to select events with a price between 55 and 60 (inclusive).

select * from StockTickEvent where price between 55 and 60

The equivalent expression without between is:

select * from StockTickEvent where price >= 55 and price <= 60

And also equivalent to:

select * from StockTickEvent where price between 60 and 55

7.9. The 'like' Keyword

The like keyword provides standard SQL pattern matching. SQL pattern matching allows you to use '_' to match any single character and '%' to match an arbitrary number of characters (including zero characters). In Esper, SQL patterns are case-sensitive by default. The syntax of like is:

test_expression [not] like pattern_expression [escape string_literal]

The test_expression is any valid expression yielding a String-type or a numeric result. The optional not keyword specifies that the result of the predicate be negated. The like keyword is followed by any valid standard SQL pattern_expression yielding a String-typed result. The optional escape keyword signals the escape character to escape '_' and '%' values in the pattern.

The result of a like expression is of type Boolean. If the value of test_expression matches the pattern_expression, the result value is true. Otherwise, the result value is false.

An example for the like keyword is below.

select * from PersonLocationEvent where name like '%Jack%'

The escape character can be defined as follows. In this example the where-clause matches events where the suffix property is a single '_' character.

select * from PersonLocationEvent where suffix like '!_' escape '!'

7.10. The 'regexp' Keyword

The regexp keyword is a form of pattern matching based on regular expressions implemented through the System.Text.RegularExpression provider. The syntax of regexp is:

test_expression [not] regexp pattern_expression

The test_expression is any valid expression yielding a String-type or a numeric result. The optional not keyword specifies that the result of the predicate be negated. The regexp keyword is followed by any valid regular expression pattern_expression yielding a String-typed result.

The result of a regexp expression is of type Boolean. If the value of test_expression matches the regular expression pattern_expression, the result value is true. Otherwise, the result value is false.

An example for the regexp keyword is below.

select * from PersonLocationEvent where name regexp '*Jack*'

7.11. The 'any' and 'some' Keywords

The any operator is true if the expression returns true for one or more of the values returned by a list of expressions including array, Collection and Map values.

The synopsis for the any keyword is as follows:

expression operator any (expression [,expression...] )

The left-hand expression is evaluated and compared to each expression result using the given operator, which must yield a Boolean result. The result of any is "true" if any true result is obtained. The result is "false" if no true result is found (including the special case where the expressions are collections that return no rows).

The operator can be any of the following values: =, !=, <>, <, <=, >, >=.

The some keyword is a synonym for any. The in construct is equivalent to = any.

Expression may also return an array, a System.Collections.Generic.ICollection or a System.Collections.Generic.IDictionary. Thus event properties that are lists, sets or maps may provide values to compare against.

All expressions must be of the same type or a compatible type. The any keyword coerces number values to compatible types. If expression returns an array, then the component type of the array must be compatible, unless the component type of the array is Object.

If expression returns an array, the operation compares each element of the array.

If expression returns a Collection, the operation determines if the collection contains the value returned by the left-hand expression, applying contains semantics. When using relationship operators <, <=, >, >= the operator applies to each element in the collection, and non-numeric elements are ignored.

If expression returns a Map, the operation determines if the map contains the key value returned by the left-hand expression, applying containsKey semantics. When using relationship operators <, <=, >, >= the operator applies to each key in the map, and non-numeric map keys are ignored.

Constants, arrays, Collection and Map expressions or event properties can be used combined.

The next statement demonstrates the use of the any operator:

select * from ProductOrder where category != any (categoryArray)

The above query selects ProductOrder event that have a category field and a category array, and returns only those events in which the category value is not in the array.

Note that if there are no successes and at least one right-hand row yields null for the operator's result, the result of the any construct will be null, not false. This is in accordance with SQL's normal rules for Boolean combinations of null values.

7.12. The 'all' Keyword

The all operator is true if the expression returns true for all of the values returned by a list of expressions including array, Collection and Map values.

The synopsis for the all keyword is as follows:

expression operator all (expression [,expression...] )

The left-hand expression is evaluated and compared to each expression result using the given operator, which must yield a Boolean result. The result of all is "true" if all rows yield true (including the special case where the expressions are collections that returns no rows). The result is "false" if any false result is found. The result is null if the comparison does not return false for any row, and it returns null for at least one row.

The operator can be any of the following values: =, !=, <>, <, <=, >, >=.

The not in construct is equivalent to != all.

Expression may also return an array, a System.Collection.Generic.ICollection or a System.Collection.Generic.IDictionary. Thus event properties that are lists, sets or maps may provide values to compare against.

All expressions must be of the same type or a compatible type. The all keyword coerces number values to compatible types. If expression returns an array, then the component type of the array must be compatible, unless the component type of the array is Object.

If expression returns an array, the operation compares each element of the array.

If expression returns a Collection, the operation determines if the collection contains the value returned by the left-hand expression, applying contains semantics. When using relationship operators <, <=, >, >= the operator applies to each element in the collection, and non-numeric elements are ignored.

If expression returns a Map, the operation determines if the map contains the key value returned by the left-hand expression, applying containsKey semantics. When using relationship operators <, <=, >, >= the operator applies to each key in the map, and non-numeric map keys are ignored.

Constants, arrays, Collection and Map expressions or event properties can be used combined.

The next statement demonstrates the use of the all operator:

select * from ProductOrder where category = all (categoryArray)

The above query selects ProductOrder event that have a category field and a category array, and returns only those events in which the category value matches all values in the array.

Chapter 8. EPL Reference: Functions

8.1. Single-row Function Reference

Single-row functions return a single value for every single result row generated by your statement. These functions can appear anywhere where expressions are allowed.

Esper allows static library methods as single-row functions, and also features built-in single-row functions. In addition, Esper allows instance method invocations on named streams.

You may also register your own single-row function name with the engine so that your EPL statements become less cluttered. This is described in detail in Section 13.1, “Custom Single-Row Functions”. Single-row functions that return an object can be chained.

Esper auto-imports the following library namespaces:

  • System

  • System.Collections

  • System.Text

Thus static library methods can be used in all expressions as shown in below example:

select symbol, Math.Round(volume/1000)
from StockTickEvent.win:time(30 sec)

In general, arbitrary type names have to be fully qualified (e.g. System.Math) but Esper provides a mechanism for user-controlled imports of classes and packages as outlined in Section 11.4.5, “Class and package imports”.

The below table outlines the built-in single-row functions available.

Table 8.1. Syntax and results of single-row functions

Single-row FunctionResult
case value 
  when compare_value then result
  [when compare_value then result ...] 
  [else result] 
  end 

Returns result where the first value equals compare_value.

case 
  when condition then result
  [when condition then result ...] 
  [else result] 
  end

Returns the result for the first condition that is true.

cast(expression, type_name)

Casts the result of an expression to the given type.

coalesce(expression, expression [, expression ...])

Returns the first non-null value in the list, or null if there are no non-null values.

current_timestamp[()]

Returns the current engine time as a long millisecond value. Reserved keyword with optional parenthesis.

exists(dynamic_property_name)

Returns true if the dynamic property exists for the event, or false if the property does not exist.

instanceof(expression, type_name [, type_name ...])

Returns true if the expression returns an object whose type is one of the types listed.

max(expression, expression [, expression ...])

Returns the highest numeric value among the 2 or more comma-separated expressions.

min(expression, expression [, expression ...])

Returns the lowest numeric value among the 2 or more comma-separated expressions.

prev(expression, event_property)

Returns a property value or all properties of a previous event, relative to the event order within a data window, or according to an optional index parameter (N) the positional Nth-from-last value.

prevtail(expression, event_property)

Returns a property value or all properties of the first event in a data window relative to the event order within a data window, or according to an optional index parameter (N) the positional Nth-from-first value.

prevwindow(event_property)

Returns a single property value of all events or all properties of all events in a data window in the order that reflects the sort order of the data window.

prevcount(event_property)

Returns the count of events (number of data points) in a data window.

prior(integer, event_property)

Returns a property value of a prior event, relative to the natural order of arrival of events

typeof(expression)

If expression is a stream name, returns the event type name of the evaluated event, often used with variant streams. If expression is a property name or expression, returns the name of the expression result type.

8.1.1. The Case Control Flow Function

The case control flow function has two versions. The first version takes a value and a list of compare values to compare against, and returns the result where the first value equals the compare value. The second version takes a list of conditions and returns the result for the first condition that is true.

The return type of a case expression is the compatible aggregated type of all return values.

The example below shows the first version of a case statement. It has a String return type and returns the value 'one'.

select case 1 when 1 then 'one' when 2 then 'two' else 'more' end from ...

The second version of the case function takes a list of conditions. The next example has a Boolean return type and returns the boolean value true.

select case when 1>0 then true else false end from ...

8.1.2. The Cast Function

The cast function casts the return type of an expression to a designated type. The function accepts two parameters: The first parameter is the property name or expression that returns the value to be casted. The second parameter is the type to cast to.

Valid parameters for the second (type) parameter are:

  • Any of the built-in types: int, long, ulong, byte, sbyte, short, ushort, char, decimal, double, float, string, where string is a short notation for System.String. The type name is not case-sensitive. For example:

    cast(price, double)

  • The fully-qualified type name of the class to cast to, for example:

    cast(product, org.myproducer.Product)

The cast function is often used to provide a type for dynamic (unchecked) properties. Dynamic properties are properties whose type is not known at compile type. These properties are always of type System.Object.

The cast function as shown in the next statement casts the dynamic "price" property of an "item" in the OrderEvent to a double value.

select cast(item.price?, double) from OrderEvent

The cast function returns a null value if the expression result cannot be casted to the desired type, or if the expression result itself is null.

The cast function adheres to the following type conversion rules:

  • For all numeric types, the cast function utilizes an internal cast conversion mecuanism to convert numeric types, if required.

  • For casts to string or System.String, the function calls System.Convert.ToString on the expression result.

  • For casts to other objects including application objects, the cast function considers a type's superclasses as well as all directly or indirectly-implemented interfaces by superclasses .

8.1.3. The Coalesce Function

The result of the coalesce function is the first expression in a list of expressions that returns a non-null value. The return type is the compatible aggregated type of all return values.

This example returns a String-typed result of value 'foo':

select coalesce(null, 'foo') from ...

8.1.4. The Current_Timestamp Function

The current_timestamp function is a reserved keyword and requires no parameters. The result of the current_timestamp function is the long-type millisecond value of the current engine system time.

The function returns the current engine timestamp at the time of expression evaluation. When using external-timer events, the function provides the last value of the externally-supplied time at the time of expression evaluation.

This example selects the current engine time:

select current_timestamp from MyEvent
// equivalent to
select current_timestamp() from MyEvent

8.1.5. The Exists Function

The exists function returns a boolean value indicating whether the dynamic property, provided as a parameter to the function, exists on the event. The exists function accepts a single dynamic property name as its only parameter.

The exists function is for use with dynamic (unchecked) properties. Dynamic properties are properties whose type is not known at compile type. Dynamic properties return a null value if the dynamic property does not exists on an event, or if the dynamic property exists but the value of the dynamic property is null.

The exists function as shown next returns true if the "item" property contains an object that has a "serviceName" property. It returns false if the "item" property is null, or if the "item" property does not contain an object that has a property named "serviceName" :

select exists(item.serviceName?) from OrderEvent

8.1.6. The Instance-Of Function

The instanceof function returns a boolean value indicating whether the type of value returned by the expression is one of the given types. The first parameter to the instanceof function is an expression to evaluate. The second and subsequent parameters are type names.

The function determines the return type of the expression at runtime by evaluating the expression, and compares the type of object returned by the expression to the defined types. If the type of object returned by the expression matches any of the given types, the function returns true. If the expression returned null or a type that does not match any of the given types, the function returns false.

The instanceof function is often used in conjunction with dynamic (unchecked) properties. Dynamic properties are properties whose type is not known at compile type.

This example uses the instanceof function to select different properties based on the type:

select case when instanceof(item, com.mycompany.Service) then serviceName?
  when instanceof(item, com.mycompany.Product) then productName? end 
  from OrderEvent

The instanceof function returns false if the expression tested by instanceof returned null.

Valid parameters for the type parameter list are:

  • Any of the built-in types: int, long, byte, short, char, double, float, string, where string is a short notation for System.String. The type name is not case-sensitive. For example, the next function tests if the dynamic "price" property is either of type float or type double:

    instanceof(price?, double, float)

  • The fully-qualified class name of the class to cast to, for example:

    instanceof(product, org.myproducer.Product)

The function considers an event class's superclasses as well as all the directly or indirectly-implemented interfaces by superclasses.

8.1.7. The Min and Max Functions

The min and max function take two or more parameters that itself can be expressions. The min function returns the lowest numeric value among the 2 or more comma-separated expressions, while the max function returns the highest numeric value. The return type is the compatible aggregated type of all return values.

The next example shows the max function that has a Double return type and returns the value 1.1.

select max(1, 1.1, 2 * 0.5) from ...

The min function returns the lowest value. The statement below uses the function to determine the smaller of two timestamp values.

select symbol, min(ticks.timestamp, news.timestamp) as minT
	from StockTickEvent.win:time(30 sec) as ticks, NewsEvent.win:time(30 sec) as news
	where ticks.symbol = news.symbol

8.1.8. The Previous Function

The prev function returns the property value or all event properties of a previous event. For data windows that introduce a sort order other then the order of arrival, such as the sorted data window and the time order data window, the function returns the event at the specified position.

The prev function is not an aggregation function and therefore does not return results per group when used with group by. Please consider the last, lastever or nth aggregation functions instead as described in Section 8.2.2, “Data Window Aggregation Functions”. You must use an aggregation function instead of prev when querying a named window.

The first parameter to the prev function is an index parameter and denotes the i-th previous event, in the order established by the data window. If no index is provided, the default index is 1 and the function returns the previous event. The second parameter is a property name or stream name. If specifying a property name, the function returns the value for the previous event property value. If specifying a stream name, the function returns the previous event underlying object.

This example selects the value of the price property of the 2nd-previous event from the current Trade event:

select prev(2, price) from Trade.win:length(10)

By using the stream alias in the previous function, the next example selects the trade event itself that is immediately previous to the current Trade event

select prev(1, trade) from Trade.win:length(10) as trade

Since the prev function takes the order established by the data window into account, the function works well with sorted windows.

In the following example the statement selects the symbol of the 3 Trade events that had the largest, second-largest and third-largest volume.

select prev(0, symbol), prev(1, symbol), prev(2, symbol)
  from Trade.ext:sort(3, volume desc)

The i-th previous event parameter can also be an expression returning an Integer-type value. The next statement joins the Trade data window with an RankSelectionEvent event that provides a rank property used to look up a certain position in the sorted Trade data window:

select prev(rank, symbol) from Trade.ext:sort(10, volume desc), RankSelectionEvent unidirectional

The prev function returns a null value if the data window does not currently hold the i-th previous event. The example below illustrates this using a time batch window. Here the prev function returns a null value for any events in which the previous event is not in the same batch of events. Note that the prior function as discussed below can be used if a null value is not the desired result.

select prev(1, symbol) from Trade.win:time_batch(1 min)

An alternative form of the prev function allows the index to not appear or appear after the property name if the index value is a constant and not an expression:

select prev(1, symbol) from Trade
// ... equivalent to ...
select prev(symbol) from Trade
// ... and ...
select prev(symbol, 1) from Trade

The combination of the prev function and std:groupwin view returns the property value for a previous event in the given data window group.

The following example returns for each event the current smallest price per symbol:

select symbol, prev(0, price) as topPricePerSymbol 
from Trade.std:groupwin(symbol).ext:sort(1, price asc)

8.1.8.1. Restrictions

The following restrictions apply to the prev functions and its results:

  • The function always returns a null value for remove stream (old data) events.

  • The function requires a data window view, or a std:groupwin and data window view, without any additional sub-views. See Chapter 9, EPL Reference: Views for built-in data window views.

8.1.8.2. Comparison to the prior Function

The prev function is similar to the prior function. The key differences between the two functions are as follows:

  • The prev function returns previous events in the order provided by the data window, while the prior function returns prior events in the order of arrival as posted by a stream's declared views.

  • The prev function requires a data window view while the prior function does not have any view requirements.

  • The prev function returns the previous event grouped by a criteria by combining the std:groupwin view and a data window. The prior function returns prior events posted by the last view regardless of data window grouping.

  • The prev function returns a null value for remove stream events, i.e. for events leaving a data window. The prior function does not have this restriction.

8.1.9. The Previous-Tail Function

The prevtail function returns the property value or all event properties of the positional-first event in a data window. For data windows that introduce a sort order other then the order of arrival, such as the sorted data window and the time order data window, the function returns the first event at the specified position.

The prevtail function is not an aggregation function and therefore does not return results per group when used with group by. Please consider the first, firstever or window aggregation functions instead as described in Section 8.2.2, “Data Window Aggregation Functions”. You must use an aggregation function instead of prevtail when querying a named window.

The first parameter is an index parameter and denotes the i-th from-first event in the order established by the data window. If no index is provided the default is zero and the function returns the first event in the data window. The second parameter is a property name or stream name. If specifying a property name, the function returns the value for the previous event property value. If specifying a stream name, the function returns the previous event underlying object.

This example selects the value of the price property of the first (oldest) event held in the length window:

select prevtail(price) from Trade.win:length(10)

By using the stream alias in the prevtail function, the next example selects the trade event itself that is the second event held in the length window:

select prevtail(1, trade) from Trade.win:length(10) as trade

Since the prevtail function takes the order established by the data window into account, the function works well with sorted windows.

In the following example the statement selects the symbol of the 3 Trade events that had the smallest, second-smallest and third-smallest volume.

select prevtail(0, symbol), prevtail(1, symbol), prevtail(2, symbol)
  from Trade.ext:sort(3, volume asc)

The i-th previous event parameter can also be an expression returning an Integer-type value. The next statement joins the Trade data window with an RankSelectionEvent event that provides a rank property used to look up a certain position in the sorted Trade data window:

select prevtail(rank, symbol) from Trade.ext:sort(10, volume asc), RankSelectionEvent unidirectional

The prev function returns a null value if the data window does not currently holds positional-first or the Nth-from-first event. For batch data windows the value returned is relative to the current batch.

The following example returns the first and second symbol value in the batch:

select prevtail(0, symbol), prevtail(1, symbol) from Trade.win:time_batch(1 min)

An alternative form of the prevtail function allows the index to not appear or appear after the property name if the index value is a constant and not an expression:

select prevtail(1, symbol) from Trade
// ... equivalent to ...
select prevtail(symbol) from Trade
// ... and ...
select prevtail(symbol, 1) from Trade

The combination of the prevtail function and std:groupwin view returns the property value for a positional first event in the given data window group.

Let's look at an example. This statement outputs the oldest price per symbol retaining the last 10 prices per symbol:

select symbol, prevtail(0, price) as oldestPrice
from Trade.std:groupwin(symbol).win:length(10)

8.1.9.1. Restrictions

The following restrictions apply to the prev functions and its results:

  • The function always returns a null value for remove stream (old data) events.

  • The function requires a data window view, or a std:groupwin and data window view, without any additional sub-views. See Chapter 9, EPL Reference: Views for built-in data window views.

8.1.10. The Previous-Window Function

The prevwindow function returns property values or all event properties for all events in a data window. For data windows that introduce a sort order other then the order of arrival, such as the sorted data window and the time order data window, the function returns the event data sorted in that order, otherwise it returns the events sorted by order of arrival with the newest arriving event first.

The prevwindow function is not an aggregation function and therefore does not return results per group when used with group by. Please consider the window aggregation function instead as described in Section 8.2.2, “Data Window Aggregation Functions”. You must use an aggregation function instead of prevwindow when querying a named window.

The single parameter is a property name or stream name. If specifying a property name, the function returns the value of the event property for all events held by the data window. If specifying a stream name, the function returns the event underlying object for all events held by the data window.

This example selects the value of the price property of all events held in the length window:

select prevwindow(price) from Trade.win:length(10)

By using the stream alias in the prevwindow function, the next example selects all trade events held in the length window:

select prevwindow(trade) from Trade.win:length(10) as trade

When used with a data window that introduces a certain sort order, the prevwindow function returns events sorted according to that sort order.

The next statement outputs for every arriving event the current 10 underying trade event objects that have the largest volume:

select prevwindow(trade) from Trade.ext:sort(10, volume desc) as trade

The prevwindow function returns a null value if the data window does not currently hold any events.

The combination of the prevwindow function and std:groupwin view returns the property value(s) for all events in the given data window group.

This example statement outputs all prices per symbol retaining the last 10 prices per symbol:

select symbol, prevwindow(price) from Trade.std:groupwin(symbol).win:length(10)

8.1.10.1. Restrictions

The following restrictions apply to the prev functions and its results:

  • The function always returns a null value for remove stream (old data) events.

  • The function requires a data window view, or a std:groupwin and data window view, without any additional sub-views. See Chapter 9, EPL Reference: Views for built-in data window views.

8.1.11. The Previous-Count Function

The prevcount function returns the number of events held in a data window.

The prevcount function is not an aggregation function and therefore does not return results per group when used with group by. Please consider the count(*) aggregation function instead as described in Section 8.2, “Aggregate Functions”. You must use an aggregation function instead of prevcount when querying a named window.

The single parameter is a property name or stream name of the data window to return the count for.

This example selects the number of data points for the price property held in the length window:

select prevcount(price) from Trade.win:length(10)

By using the stream alias in the prevcount function the next example selects the count of trade events held in the length window:

select prevcount(trade) from Trade.win:length(10) as trade

The combination of the prevcount function and std:groupwin view returns the count of events in the given data window group.

This example statement outputs the number of events retaining the last 10 events per symbol:

select symbol, prevcount(price) from Trade.std:groupwin(symbol).win:length(10)

8.1.11.1. Restrictions

The following restrictions apply to the prev functions and its results:

  • The function always returns a null value for remove stream (old data) events.

  • The function requires a data window view, or a std:groupwin and data window view, without any additional sub-views. See Chapter 9, EPL Reference: Views for built-in data window views.

8.1.12. The Prior Function

The prior function returns the property value of a prior event. The first parameter is an integer value that denotes the i-th prior event in the natural order of arrival. The second parameter is a property name for which the function returns the value for the prior event. The second parameter is a property name or stream name. If specifying a property name, the function returns the property value for the prior event. If specifying a stream name, the function returns the prior event underlying object.

This example selects the value of the price property of the 2nd-prior event to the current Trade event.

select prior(2, price) from Trade

By using the stream alias in the prior function, the next example selects the trade event itself that is immediately prior to the current Trade event

select prior(1, trade) from Trade as trade

The prior function can be used on any event stream or view and does not have any specific view requirements. The function operates on the order of arrival of events by the event stream or view that provides the events.

The next statement uses a time batch window to compute an average volume for 1 minute of Trade events, posting results every minute. The select-clause employs the prior function to select the current average and the average before the current average:

select average, prior(1, average) 
    from TradeAverages.win:time_batch(1 min).stat:uni(volume)

8.1.13. The Type-Of Function

The typeof function, when parameterized by a stream name, returns the event type name of the evaluated event which can be useful with variant streams. When parameterized by an expression or property name, the function returns the type name of the expression result or null if the expression result is null.

In summary, the function determines the return type of the expression at runtime by evaluating the expression and returns the type name of the expression result.

The typeof function is often used in conjunction with variant streams. A variant stream is a predefined stream into which events of multiple disparate event types can be inserted. The typeof function, when passed a stream name alias, returns the name of the event type of each event in the stream.

The following example elaborates on the use of variant streams with typeof. The first statement declares a variant stream SequencePatternStream:

create variant schema SequencePatternStream as *

The next statement inserts all order events and is followed by a statement to insert all product events:

insert into SequencePatternStream select * from OrderEvent;
insert into SequencePatternStream select * from PriceEvent;

This example statement returns the event type name for each event in the variant stream:

select typeof(sps) from SequencePatternStream as sps

The next example statement detects a pattern by utilizing the typeof function to find pairs of order event immediately followed by product event:

select * from SequencePatternStream match_recognize(
  measures A as a, B as b
  pattern (A B)
  define A as typeof(A) = "OrderEvent",
         B as typeof(B) = "ProductEvent"
  )

When passing a property name to the typeof function, the function evaluates whether the property type is event type (a fragment event type). If the property type is event type, the function returns the type name of the event in the property value or null if not provided. If the property type is not event type, the function returns the simple class name of the property value.

When passing an expression to the typeof function, the function evaluates the expression and returns the simple class name of the expression result value or null if the expression result value is null.

This example statement returns the simple class name of the value of the dynamic property prop of events in stream MyStream, or a null value if the property is not found for an event or the property value itself is null:

select typeof(prop?) from MyStream

When using subclasses or interface implementations as event classes or when using Map-event type inheritance, the function returns the event type name provided when the class or Map-type event was registered, or if the event type was not registered, the function returns the fully-qualified class name.

8.2. Aggregate Functions

Aggregation functions return a single value from a collection of input values. The group by keywords are often used in conjunction with aggregation functions to group the result-set by one or more columns.

The EPL language extends the standard SQL aggregation functions by useful aggregation functions that can track a data window or compute event rates. Your application may also add its own aggregation function as Section 13.3, “Custom Aggregation Functions” describes.

Aggregation values are always computed incrementally: Insert and remove streams result in aggregation value changes. The exceptions are on-demand queries and joins when using the unidirectional keyword. Aggregation functions are optimized to retain the minimal information necessary to compute the aggregated result.

Aggregation functions can also be used with unbound streams when no data window is specified.

8.2.1. SQL-Standard Functions

The SQL-standard aggregation functions are shown in below table.

Table 8.2. Syntax and results of SQL-standard aggregation functions

Aggregate FunctionResult
avedev([all|distinct] expression)

Mean deviation of the (distinct) values in the expression, returning a value of double type.

avg([all|distinct] expression)

Average of the (distinct) values in the expression, returning a value of double type.

count([all|distinct] expression)

Number of the (distinct) non-null values in the expression, returning a value of long type.

count(*)

Number of events, returning a value of long type.

max([all|distinct] expression)

Highest (distinct) value in the expression, returning a value of the same type as the expression itself returns.

median([all|distinct] expression)

Median (distinct) value in the expression, returning a value of double type. Double Not-a-Number (NaN) values are ignored in the median computation.

min([all|distinct] expression)

Lowest (distinct) value in the expression, returning a value of the same type as the expression itself returns.

stddev([all|distinct] expression)

Standard deviation of the (distinct) values in the expression, returning a value of double type.

sum([all|distinct] expression)

Totals the (distinct) values in the expression, returning a value of long, double, float or integer type depending on the expression.

If your application provides double-type values to an aggregation function, avoid using Not-a-Number (NaN) and infinity. Also when using double-type values, round-off errors (or rounding errors) may occur due to double-type precision. Consider rounding your result value to the desired precision.

8.2.2. Data Window Aggregation Functions

The first, last and window aggregation functions return event properties of events present in a stream's data window. They are useful when information about current data window contents is required.

When comparing the last aggregation function to the prev function, the differences are as follows. The prev function is not an aggregation function and thereby not sensitive to the presence of group by. The prev function accesses data window contents directly and respects the sort order of the data window. The last aggregation function returns results based on arrival order and tracks data window contents in a separate shared data structure.

When comparing the first aggregation function to the prevtail function, the differences are as follows. The prevtail function is not an aggregation function and thereby not sensitive to the presence of group by. The prevtail function accesses data window contents directly and respects the sort order of the data window. The first aggregation function returns results based on arrival order and tracks data window contents in a separate shared data structure.

When comparing the window aggregation function to the prevwindow function, the differences are as follows. The prevwindow function is not an aggregation function and thereby not sensitive to the presence of group by. The prevwindow function accesses data window contents directly and respects the sort order of the data window. The window aggregation function returns results based on arrival order and tracks data window contents in a separate shared data structure.

When comparing the count aggregation function to the prevcount function, the differences are as follows. The prevcount function is not an aggregation function and thereby not sensitive to the presence of group by.

When comparing the last aggregation function to the nth aggregation function, the differences are as follows. The nth aggregation function does not consider out-of-order deletes (for example with on-delete and sorted windows) and does not revert to the prior expression value when the last event or nth-event was deleted from a data window. The last aggregation function tracks the data window and reflects out-of-order deletes.

From an implementation perspective, the first, last and window aggregation functions share a common data structure for each stream.

8.2.2.1. First Aggregation Function

The synopsis for the first aggregation function is:

first(*|stream.*|value_expression [, index_expression])

The first aggregation function returns properties of very first event in the data window. When used with group by, it returns properties of the first event in the data window for each group. When specifying an index expression, the function returns properties of the Nth-subsequent event to the first event, all according to order of arrival.

The first parameter to the function is required and defines the event properties or expression result to return. The second parameter is an optional index_expression that must return an integer value used as an index to evaluate the Nth-subsequent event to the first event.

You may specify the wildcard (*) character in which case the function returns the underlying event of the single selected stream. For joins and subqueries you must use the stream wildcard syntax below.

You may specify the stream name and wildcard (*) character in the stream.* syntax. This returns the underlying event for the specified stream.

You may specify a value_expression to evaluate for the first event. The value expression may not select properties from multiple streams.

The index_expression is optional. If no index expression is provided, the function returns the first event. If present, the function evaluates the index expression to determine the value for N, and evaluates the Nth-subsequent event to the first event. A value of zero returns the first event and a value of 1 returns the event subsequent to the first event. You may not specify event properties in the index expression.

The function returns null if there are no events in the data window or when the index is larger then the number of events held in the data window. If used with group by, it returns null if there are no events in the data window for that group or when the index is larger then the number of events held in the data window for that group.

To explain, consider the statement below which selects the underlying event of the first sensor event held by the length window of 2 events.

select first(*) from SensorEvent.win:length(2) 

Assume event E1, event E2 and event E3 are of type SensorEvent. When event E1 arrives the statement outputs the underlying event E1. When event E2 arrives the statement again outputs the underlying event E1. When event E3 arrives the statement outputs the underlying event E2, since event E1 has left the data window.

The stream wildcard syntax is useful for joins and subqueries. This example demonstrates a subquery that returns the first SensorEvent when a DoorEvent arrives:

select (select first(se.*) from SensorEvent.win:length(2) as se) from DoorEvent 

The following example shows the use of an index expression. The output value for f1 is the temperature property value of the first event in the data window, the value for f2 is the temperature property value of the second event in the data window:

select first(temperature, 0) as f1, first(temperature, 1) as f2
from SensorEvent.win:time(10 sec)

8.2.2.2. Last Aggregation Function

The synopsis for the last aggregation function is:

last(*|stream.*|value_expression [, index_expression])

The last aggregation function returns properties of the very last event in the data window. When used with group by, it returns properties of the last event in the data window for each group. When specifying an index expression, the function returns properties of the Nth-prior event to the last event, all according to order of arrival.

Similar to the first aggregation function described above, you may specify the wildcard (*) character or stream name and wildcard (*) character or a value_expression to evaluate for the last event.

The index_expression is optional. If no index expression is provided, the function returns the last event. If present, the function evaluates the index expression to determine the value for N, and evaluates the Nth-prior event to the last event. A value of zero returns the last event and a value of 1 returns the event prior to the last event. You may not specify event properties in the index expression.

The function returns null if there are no events in the data window or when the index is larger then the number of events held in the data window. If used with group by, it returns null if there are no events in the data window for that group or when the index is larger then the number of events held in the data window for that group.

The next statement selects the underlying event of the first and last sensor event held by the time window of 10 seconds:

select first(*), last(*) from SensorEvent.win:time(10 sec) 

The statement shown next selects the last temperature (f1) and the prior-to-last temperature (f1) of sensor events in the last 10 seconds:

select last(temperature, 0) as f1, select last(temperature, 1) as f2
from SensorEvent.win:time(10 sec)

8.2.2.3. Window Aggregation Function

The synopsis for the window aggregation function is:

window(*|stream.*|value_expression)

The window aggregation function returns properties of all events in the data window. When used with group by, it returns properties of all events in the data window for each group.

Similar to the first aggregation function described above, you may specify the wildcard (*) character or stream name and wildcard (*) character or a value_expression to evaluate for all events.

The function returns null if there are no events in the data window. If used with group by, it returns null if there are no events in the data window for that group.

The next statement selects the underlying event of all events held by the time window of 10 seconds:

select window(*) from SensorEvent.win:time(10 sec) 

The window aggregation function requires that your stream is bound by a data window. You may not use the window aggregation function on unbound streams with the exception of on-demand queries.

8.2.3. Additional Aggregation Functions

Esper provides the following additional aggregation functions beyond those in the SQL standard:

Table 8.3. Syntax and results of EPL aggregation functions

Aggregate FunctionResult
firstever(expression)

The firstever aggregation function returns the very first value ever. If used with group by it returns the first value ever for that group.

If used with a data window, the result of the function does not change as data points leave a data window. Use the first or prevtail function to return values relative to a data window.

lastever(expression)

Returns the last value or last value per group, if used with group by.

This sample statement outputs the total price, the first price and the last price per symbol for the last 30 seconds of events and every 5 seconds:

select symbol, sum(price), lastever(price), firstever(price)
from StockTickEvent.win:time(30 sec) 
group by symbol
output every 5 sec

If used with a data window, the result of the function does not change as data points leave a data window (for example when all data points leave the data window). Use the last or prev function to return values relative to a data window.

leaving()

Returns true when any remove stream data has passed, for use in the having clause to output only when a data window has filled.

The leaving aggregation function is useful when you want to trigger output after a data window has a remove stream data point. Use the output after syntax as an alternative to output after a time interval.

This sample statement uses leaving() to output after the first data point leaves the data window, ignoring the first datapoint:

select symbol, sum(price) 
from StockTickEvent.win:time(30 sec) 
having leaving()
nth(expression, N_index)

Returns the Nth oldest element; If N=0 returns the most recent value. If N=1 returns the value before the most recent value. If N is larger than the events held in the data window for this group, returns null.

A maximum N historical values are stored, so it can be safely used to compare recent values in large views without incurring excessive overhead.

As compared to the prev row function, this aggregation function works within the current group by group, see Section 3.7.2, “Output for Aggregation and Group-By”.

This statement outputs every 2 seconds the groups that have new data and their last price and the previous-to-last price:

select symbol, nth(price, 1), last(price) 
from StockTickEvent 
group by symbol
output last every 2 sec

rate(number_of_seconds)

Returns an event arrival rate per second over the provided number of seconds, computed based on engine time.

Returns null until events fill the number of seconds. Useful with output snapshot to output a current rate. Does not require a data window onto the stream(s).

A sample statement to output, every 2 seconds, the arrival rate per second considering the last 10 seconds of events is shown here:

select rate(10) from StockTickEvent
output snapshot every 2 sec

The aggregation function retains an engine timestamp value for each arriving event.

rate(timestamp_property[, accumulator])

Returns an event arrival rate over the data window including the last remove stream event. The timestamp_property is the name of a long-type property of the event that provides a timestamp value.

The first parameter is a property name or expression providing millisecond timestamp values.

The optional second parameter is a property or expression for computing an accumulation rate: If a value is provided as a second parameter then the accumulation rate for that quantity is returned (e.g. turnover in dollars per second).

Requires a data window declared onto the stream. Returns null until events start leaving the window.

This sample statement outputs event rate for each group (symbol) with fixed sample size of four events (and considering the last event that left). The timestamp event property must be part of the event for this to work.

select colour, rate(timestamp) as rate 
from StockTickEvent.std:groupwin(symbol).win:length(4) 
group by symbol

Built-in aggregation functions can be disabled via configuration (see Section 11.4.18.4, “Extended Built-in Aggregation Functions”). A custom aggregation function of the same name as a built-on function may be registered to override the built-in function.

8.3. User-Defined Functions

A user-defined function (UDF) is a single-row function that can be invoked anywhere as an expression itself or within an expresson. The function must simply be a public static method that the classloader can resolve at statement creation time. The engine resolves the function reference at statement creation time and verifies parameter types.

You may register your own single-row function name for the user-defined function so that your EPL statements are less cluttered. This is described in detail in Section 13.1, “Custom Single-Row Functions”.

User-defined functions can be also be invoked on instances of an event: Please see Section 4.4.5, “Using the Stream Name” to invoke event instance methods on a named stream.

The example below assumes a class MyClass that exposes a public static method myFunction accepting 2 parameters, and returing a numeric type such as double.

select 3 * com.mycompany.MyClass.MyFunction(price, volume) as myValue 
from StockTick.win:time(30 sec)

User-defined functions also take array parameters as this example shows. The section on Section 7.5, “Array Definition Operator” outlines in more detail the types of arrays produced.

select * from RFIDEvent where com.mycompany.rfid.MyChecker.IsInZone(zone, {10, 20, 30})

Type names have to be fully qualified (e.g. System.Math) but Esper provides a mechanism for user-controlled imports of classes and packages as outlined in Section 11.4.5, “Class and package imports”.

User-defined functions can return any value including null, native objects or arrays. Therefore user-defined functions can serve to transform, convert or map events, or to extract information and assemble further events.

The following statement is a simple pattern that looks for events of type E1 that are followed by events of type E2. It assigns the tags "e1" and "e2" that the function can use to assemble a final event for output:

select MyLib.MapEvents(e1, e2) from pattern [every e1=E1 -> e2=E2]

User-defined functions may also be chained: If a user-defined function returns an object then the object can itself be the target of the next function call and so on.

Assume that there is a calculator function in the MyLib class that returns a class which provides the Search method taking two parameters. The EPL that takes the result of the Calculator property and that calls the Search method on the result and returns its return value is shown below:

select MyLib.Calculator.Search(zonevariable, zone) from RFIDEvent]

A user-defined function should be implemented thread-safe.

8.3.1. Event Type Conversion via User-Defined Function

A function that converts from one event type to another event type is shown in the next example. The first statement declares a stream that consists of MyEvent events. The second statement employs a conversion function to convert MyOtherEvent events to events of type MyEvent:

insert into MyStream select * from MyEvent
	insert into MyStream select MyLib.Convert(other) from MyOtherEvent as other

In the example above, assuming the event classes MyEvent and MyOtherEvent are types, the static method should have the following footprint:

public static MyEvent Convert(MyOtherEvent otherEvent)

8.3.2. User-Defined Function Result Cache

For user-defined functions that take no parameters or only constants as parameters the engine automatically caches the return result of the function, and invokes the function only once. This is beneficial to performance if your function indeed returns the same result for the same input parameters.

You may disable caching of return values of user-defined functions via configuration as described in Section 11.4.18.3, “User-Defined Function or Static Method Cache”.

8.3.3. Parameter Matching

EPL follows CLR standards in terms of widening, performing widening automatically in cases where widening type conversion is allowed without loss of precision, for both boxed and primitive types.

When user-defined functions are overloaded, the function with the best match is selected based on how well the arguments to a function can match up with the parameters, giving preference to the function that requires the least number of widening conversions.

Boxing and unboxing of arrays is not supported in UDF. For example, an array of int? and an array of int are not compatible types.

When using {} array syntax in EPL, the resulting type is always a boxed type: "{1, 2}" is an array of Integer (and not int since it may contain null values), "{1.0, 2d}" is an array of double and "{'A', "B"}" is an array of string, while "{1, "B", 2.0}" is an array of Object (Object[]).

Chapter 9. EPL Reference: Views

This chapter outlines the views that are built into Esper. All views can be arbitrarily combined as many of the examples below show. The section on Chapter 3, Processing Model provides additional information on the relationship of views, filtering and aggregation. Please also see Section 4.4.3, “Specifying Views” for the use of views in the from clause with streams, patterns and named windows.

Esper organizes built-in views in namespaces and names. Views that provide sliding or tumbling data windows are in the win namespace. Other most commonly used views are in the std namespace. The ext namespace are views that order events. The stat namespace is used for views that derive statistical data.

Esper distinguishes between data window views and derived-value views. Data windows, or data window views, are views that retain incoming events until an expiry policy indicates to release events. Derived-value views derive a new value from event streams and post the result as events of a new type.

Two or more data window views can be combined. This allows a sets of events retained by one data window to be placed into a union or an intersection with the set of events retained by one or more other data windows. Please see Section 4.4.4, “Multiple Data Window Views” for more detail.

The keep-all data window counts as a data window but has no expiry policy: it retains all events received. The grouped-window declaration allocates a new data window per grouping criteria and thereby counts as a data window, but cannot appear alone.

The next table summarizes data window views:

Table 9.1. Built-in Data Window Views

ViewSyntaxDescription
Length windowwin:length(size)Sliding length window extending the specified number of elements into the past.
Length batch windowwin:length_batch(size)Tumbling window that batches events and releases them when a given minimum number of events has been collected.
Time windowwin:time(time period)Sliding time window extending the specified time interval into the past.
Externally-timed windowwin:ext_timed(timestamp expression, time period)Sliding time window, based on the millisecond time value supplied by an expression.
Time batch windowwin:time_batch(time period[,optional reference point] [, flow control])Tumbling window that batches events and releases them every specified time interval, with flow control options.
Time-Length combination batch windowwin:time_length_batch(time period, size [, flow control])Tumbling multi-policy time and length batch window with flow control options.
Time-Accumulating windowwin:time_accum(time period)Sliding time window accumulates events until no more events arrive within a given time interval.
Keep-All windowwin:keepall()The keep-all data window view simply retains all events.
Sorted windowext:sort(size, sort criteria)Sorts by values returned by sort criteria expressions and keeps only the top events up to the given size.
Time-Order Viewext:time_order(timestamp expression, time period)Orders events that arrive out-of-order, using an expression providing timestamps to be ordered.
Uniquestd:unique(unique criteria(s))Retains only the most recent among events having the same value for the criteria expression(s). Acts as a length window of size 1 for each distinct expression value.
Grouped Data Windowstd:groupwin(grouping criteria(s))Groups events into sub-views by the value of the specified expression(s), generally used to provide a separate data window per group.
Last Eventstd:lastevent()Retains the last event, acts as a length window of size 1.
First Eventstd:firstevent()Retains the very first arriving event, disregarding all subsequent events.
First Uniquestd:firstunique(unique criteria(s))Retains only the very first among events having the same value for the criteria expression(s), disregarding all subsequent events for same value(s).
First Lengthwin:firstlength(size)Retains the first size events, disregarding all subsequent events.
First Timewin:firsttime(time period)Retains the events arriving until the time interval has passed, disregarding all subsequent events.

The table below summarizes views that derive information from received events and present the derived information as an insert and remove stream of events that are typed specifically to carry the result of the computations:

Table 9.2. Built-in Derived-Value Views

ViewSyntaxDescription
Sizestd:size([expression, ...])Derives a count of the number of events in a data window, or in an insert stream if used without a data window, and optionally provides additional event properties as listed in parameters.
Univariate statisticsstat:uni(value expression [,expression, ...])Calculates univariate statistics on the values returned by the expression.
Regressionstat:linest(value expression, value expression [,expression, ...])Calculates regression on the values returned by two expressions.
Correlationstat:correl(value expression, value expression [,expression, ...])Calculates the correlation value on the values returned by two expressions.
Weighted averagestat:weighted_avg(value expression, value expression [,expression, ...])Calculates weighted average given a weight expression and an expression to compute the average for.

9.1. A Note on View Parameters

The syntax for view specifications starts with the namespace name and the name and is followed by optional view parameter expressions in parenthesis:

namespace:name(view_parameters)

This example specifies a time window of 5 seconds:

select * from StockTickEvent.win:time(5 sec)

All expressions are allowed as parameters to views, including expressions that contain variables or substitution parameters for prepared statements.

For example, assuming a variable by name VAR_WINDOW_SIZE is defined:

select * from StockTickEvent.win:time(VAR_WINDOW_SIZE)

Expression parameters for views are evaluated at the time the view is first created. They are not continuously re-evaluated by built-in views. For applications that provide a custom plug-in view, such custom views may re-evaluate parameter expressions.

If a view takes no parameters, use empty parenthesis ().

9.1. Window views

All the views explained below are data window views, as are std:unique, std:firstunique, std:lastevent and std:firstevent.

9.1.1. Length window (win:length)

This view is a moving (sliding) length window extending the specified number of elements into the past. The view takes a single expression as a parameter providing a numeric size value that defines the window size:

win:length(size_expression)

The below example sums the price for the last 5 stock ticks for symbol GE.

select sum(price) from StockTickEvent(symbol='GE').win:length(5)

The next example keeps a length window of 10 events of stock trade events, with a separate window for each symbol. The sum of price is calculated only for the last 10 events for each symbol and aggregates per symbol:

select sum(price) from StockTickEvent.std:groupwin(symbol).win:length(10) group by symbol

9.1.2. Length batch window (win:length_batch)

This window view buffers events (tumbling window) and releases them when a given minimum number of events has been collected. Provide an expression defining the number of events to batch as a parameter:

win:length_batch(size_expression)

The next statement buffers events until a minimum of 10 events have collected. Listeners to updates posted by this view receive updated information only when 10 or more events have collected.

select * from StockTickEvent.win:length_batch(10)

9.1.3. Time window (win:time)

This view is a moving (sliding) time window extending the specified time interval into the past based on the system time. Provide a time period (see Section 4.2.1, “Specifying Time Periods”) or an expression defining the number of seconds as a parameter:

win:time(time period)
win:time(seconds_interval_expression)

For the GE stock tick events in the last 1 second, calculate a sum of price.

select sum(price) from StockTickEvent(symbol='GE').win:time(1 sec)

The following time windows are equivalent specifications:

win:time(2 minutes 5 seconds)
win:time(125 sec)
win:time(125)
win:time(MYINTERVAL)  // MYINTERVAL defined as a variable

9.1.4. Externally-timed window (win:ext_timed)

Similar to the time window, this view is a moving (sliding) time window extending the specified time interval into the past, but based on the millisecond time value supplied by a timestamp expression. The view takes two parameters: the expression to return long-typed timestamp values, and a time period or expression that provides a number of seconds:

win:ext_timed(timestamp_expression, time_period)
win:ext_timed(timestamp_expression, seconds_interval_expression)

The key difference comparing the externally-timed window to the regular time window is that the window slides not based on the engine time, but strictly based on the result of the timestamp expression when evaluated against the events entering the window.

The algorithm underlying the view compares the timestamp value returned by the expression when the oldest event arrived with the timestamp value returned by the expression for the newest arriving event on event arrival. If the time interval between the timestamp values is larger then the timer period parameter, then the algorithm removes all oldest events tail-first until the difference between the oldest and newest event is within the time interval. The window therefore slides only when events arrive and only considers each event's timestamp property (or other expression value returned) and not engine time.

This view holds stock tick events of the last 10 seconds based on the timestamp property in StockTickEvent.

select * from StockTickEvent.win:ext_timed(timestamp, 10 seconds)

The externally-timed data window expects strict ordering of the timestamp values returned by the timestamp expression. The view is not useful for ordering events in time order, please us the time-order view instead.

9.1.5. Time batch window (win:time_batch)

This window view buffers events (tumbling window) and releases them every specified time interval in one update. The view takes a time period or an expression providing a number of seconds as a parameter, plus optional parameters described next.

win:time_batch(time_period [,optional_reference_point] [,flow_control])
win:time_batch(seconds_interval_expression [,optional_reference_point] [,flow_control])

The time batch window takes a second, optional parameter that serves as a reference point to batch flush times. If not specified, the arrival of the first event into the batch window sets the reference point. Therefore if the reference point is not specified and the first event arrives at time t1, then the batch flushes at time t1 plus time_period and every time_period thereafter.

The below example batches events into a 5 second window releasing new batches every 5 seconds. Listeners to updates posted by this view receive updated information only every 5 seconds.

select * from StockTickEvent.win:time_batch(5 sec)

By default, if there are no events arriving in the current interval (insert stream), and no events remain from the prior batch (remove stream), then the view does not post results to listeners. The view allows overriding this default behavior via flow control keywords.

The synopsis with flow control parameters is:

win:time_batch(time_period or seconds_interval_expr [,optional_reference_point] 
    [, "flow-control-keyword [, keyword...]"] )

The FORCE_UPDATE flow control keyword instructs the view to post an empty result set to listeners if there is no data to post for an interval. When using this keyword the irstream keyword should be used in the select clause to ensure the remove stream is also output.

The START_EAGER flow control keyword instructs the view to post empty result sets even before the first event arrives, starting a time interval at statement creation time. As when using FORCE_UPDATE, the view also posts an empty result set to listeners if there is no data to post for an interval, however it starts doing so at time of statement creation rather then at the time of arrival of the first event.

Taking the two flow control keywords in one sample statement, this example presents a view that waits for 10 seconds. It posts empty result sets after one interval after the statement is created, and keeps posting an empty result set as no events arrive during intervals:

select * from MyEvent.win:time_batch(10 sec, "FORCE_UPDATE, START_EAGER")

The optional reference point is provided as a long-value of milliseconds relative to January 1, 1970 and time 00:00:00.

The following example statement sets the reference point to 5 seconds and the batch size to 1 hour, so that each batch output is 5 seconds after each hour:

select * from OrderSummaryEvent.win:time_batch(1 hour, 5000L)

9.1.6. Time-Length combination batch window (win:time_length_batch)

This data window view is a combination of time and length batch (tumbling) windows. Similar to the time and length batch windows, this view batches events and releases the batched events when either one of the following conditions occurs, whichever occurs first: the data window has collected a given number of events, or a given time interval has passed.

The view parameters take 2 forms. The first form accepts a time period or an expression providing a number of seconds, and an expression for the number of events:

win:time_length_batch(time_period, number_of_events_expression)
win:time_length_batch(seconds_interval_expression, number_of_events_expression)

The next example shows a time-length combination batch window that batches up to 100 events or all events arriving within a 1-second time interval, whichever condition occurs first:

 select * from MyEvent.win:time_length_batch(1 sec, 100)

In this example, if 100 events arrive into the window before a 1-second time interval passes, the view posts the batch of 100 events. If less then 100 events arrive within a 1-second interval, the view posts all events that arrived within the 1-second interval at the end of the interval.

By default, if there are no events arriving in the current interval (insert stream), and no events remain from the prior batch (remove stream), then the view does not post results to listeners. This view allows overriding this default behavior via flow control keywords.

The synopsis of the view with flow control parameters is:

win:time_length_batch(time_period or seconds_interval_expression, number_of_events_expression, 
    "flow control keyword [, keyword...]")

The FORCE_UPDATE flow control keyword instructs the view to post an empty result set to listeners if there is no data to post for an interval. The view begins posting no later then after one time interval passed after the first event arrives. When using this keyword the irstream keyword should be used in the select clause to ensure the remove stream is also output.

The START_EAGER flow control keyword instructs the view to post empty result sets even before the first event arrives, starting a time interval at statement creation time. As when using FORCE_UPDATE, the view also posts an empty result set to listeners if there is no data to post for an interval, however it starts doing so at time of statement creation rather then at the time of arrival of the first event.

Taking the two flow control keywords in one sample statement, this example presents a view that waits for 10 seconds or reacts when the 5th event arrives, whichever comes first. It posts empty result sets after one interval after the statement is created, and keeps posting an empty result set as no events arrive during intervals:

 select * from MyEvent.win:time_length_batch(10 sec, 5, "FORCE_UPDATE, START_EAGER")

9.1.7. Time-Accumulating window (win:time_accum)

This data window view is a specialized moving (sliding) time window that differs from the regular time window in that it accumulates events until no more events arrive within a given time interval, and only then releases the accumulated events as a remove stream.

The view accepts a single parameter: the time period or seconds-expression specifying the length of the time interval during which no events must arrive until the view releases accumulated events. The synopsis is as follows:

win:time_accum(time_period)
win:time_accum(seconds_interval_expression)

The next example shows a time-accumulating window that accumulates events, and then releases events if within the time interval no more events arrive:

 select * from MyEvent.win:time_accum(10 sec)

This example accumulates events, until when for a period of 10 seconds no more MyEvent events arrive, at which time it posts all accumulated MyEvent events.

Your application may only be interested in the batches of events as events leave the data window. This can be done simply by selecting the remove stream of this data window, populated by the engine as accumulated events leave the data window all-at-once when no events arrive during the time interval following the time the last event arrived:

 select rstream * from MyEvent.win:time_accum(10 sec)

If there are no events arriving, then the view does not post results to listeners.

9.1.8. Keep-All window (win:keepall)

This keep-all data window view simply retains all events. The view does not remove events from the data window, unless used with a named window and the on delete clause.

The view accepts no parameters. The synopsis is as follows:

win:keepall()

The next example shows a keep-all window that accumulates all events received into the window:

 select * from MyEvent.win:keepall()

Note that since the view does not release events, care must be taken to prevent retained events from using all available memory.

9.1.9. First Length (win:firstlength)

The firstlength view retains the very first size_expression events.

The synopsis is:

win:firstlength(size_expression)

If used within a named window and an on-delete clause deletes events, the view accepts further arriving events until the number of retained events reaches the size of size_expression.

The below example creates a view that retains only the first 10 events:

select * from MyEvent.win:firstlength(10)

9.1.10. First Time (win:firsttime)

The firsttime view retains all events arriving within a given time interval after statement start.

The synopsis is:

win:firsttime(time_period)
win:firsttime(seconds_interval_expression)

The below example creates a view that retains only those events arriving within 1 minute and 10 seconds of statement start:

select * from MyEvent.win:firsttime(1 minute 10 seconds)

9.2. Standard view set

9.2.1. Unique (std:unique)

The unique view is a view that includes only the most recent among events having the same value(s) for the result of the specified expression or list of expressions.

The synopsis is:

std:unique(unique_expression [, unique_expression ...])

The view acts as a length window of size 1 for each distinct value returned by an expression, or combination of values returned by multiple expressions. It thus posts as old events the prior event of the same value(s), if any.

An expression may return a null value. The engine treats a null value as any other value. An expression can also return a custom application object, whereby the application class should implement the GetHashCode and Equals methods.

The below example creates a view that retains only the last event per symbol.

select * from StockTickEvent.std:unique(symbol)

The next example creates a view that retains the last event per symbol and feed.

select * from StockTickEvent.std:unique(symbol, feed)

9.2.2. Grouped Data Window (std:groupwin)

This view groups events into sub-views by the value returned by the specified expression or the combination of values returned by a list of expressions. The view takes a single expression to supply the group criteria values, or a list of expressions as parameters, as the synopsis shows:

std:groupwin(grouping_expression [, grouping_expression ...])

The grouping_expression expression(s) return one or more group keys, by which the view creates sub-views for each distinct group key. Note that the expression should not return an unlimited number of values: the grouping expression should not return a time value or otherwise unlimited key.

An expression may return a null value. The engine treats a null value as any other value. An expression can also return a custom application object, whereby the application class should implement the GetHashCode and Equals methods.

This example computes the total price for the last 5 events considering the last 5 events per each symbol, aggregating the price across all symbols (since no group by clause is specified the aggregation is across all symbols):

select symbol, sum(price) from StockTickEvent.std:groupwin(symbol).win:length(5)

The @Hint("reclaim_group_aged=age_in_seconds") hint instructs the engine to discard grouped data window state that has not been updated for age_in_seconds seconds. The optional @Hint("reclaim_group_freq=sweep_frequency_in_seconds") can be specified in addition to control the frequency at which the engine sweeps data window state. If the hint is not specified, the frequency defaults to the same value as age_in_seconds. Use the hints when your group criteria returns a changing or unlimited number of values. By default and without hints the view does not reclaim or remove data windows for group criteria values.

The updated sample statement with both hints:

// Remove data window views for symbols not updated for 10 seconds or more and sweep every 30 seconds
@Hint('reclaim_group_aged=10,reclaim_group_freq=30')
select symbol, sum(price) from StockTickEvent.std:groupwin(symbol).win:length(5)

To compute the total price for the last 5 events considering the last 5 events per each symbol and outputting a price per symbol, add the group by clause:

select symbol, sum(price) from StockTickEvent.std:groupwin(symbol).win:length(5) group by symbol

The std:groupwin grouped-window view can also take multiple expressions that provide values to group by. This example computes the total price for each symbol and feed for the last 10 events per symbol and feed combination:

select sum(price) from StockTickEvent.std:groupwin(symbol, feed).win:length(10)

The order in which the std:groupwin grouped-window view appears within sub-views of a stream controls the data the engine derives from events for each group. The next 2 statements demonstrate this using a length window.

Without the std:groupwin declaration query the same query returns the total price per symbol for only the last 10 events across all symbols. Here the engine allocates only one length window for all events:

select sum(price) from StockTickEvent.win:length(10)

We have learned that by placing the std:groupwin grouped-window view before other views, these other views become part of the grouped set of views. The engine dynamically allocates a new view instance for each subview, every time it encounters a new group key such as a new value for symbol. Therefore, in std:groupwin(symbol).win:length(10) the engine allocates a new length window for each distinct symbol. However in win:length(10) alone the engine maintains a single length window.

The std:groupwin can be used with multiple data window views to achieve a grouped intersection or union policy.

The next query retains the last 4 events per symbol and only those events that are also not older then 10 seconds:

select * from StockTickEvent.std:groupwin(symbol).win:length(4).win:time(10)

Last, we consider a grouped data window for two group criteria. Here, the query results are total price per symbol and feed for the last 100 events per symbol and feed.

select sum(price) from StockTickEvent.std:groupwin(symbol, feed).win:length(100)

For advanced users: There is an optional view that can control how the std:groupwin grouped-window view gets evaluated and that view is the std:merge view. The merge view can only occur after a std:groupwin grouped-window view in a view chain and controls at what point in the view chain the merge of the data stream occurs from view-instance-per-criteria to single view.

Compare the following statements:

select * from Market.std:groupwin(ticker).win:length(1000000)
    .stat:weighted_avg(price, volume).std:merge(ticker)
// ... and ...
select * from Market.std:groupwin(ticker).win:length(1000000).std:merge(ticker)
    .stat:weighted_avg(price, volume)

If your statement does not specify the optional std:merge view, the semantics are the same as the first statement.

The first statement, in which the merge-view is added to the end (same as no merge view), computes weighted average per ticker, considering, per-ticker, the last 1M Market events for each ticker. The second statement, in which the merge view is added to the middle, computes weighted average considering, per-ticker, the last 1M Market events, computing the weighted average for all such events using a single view rather then multiple view instances with one view per ticker.

9.2.3. Size (std:size)

This view posts the number of events received from a stream or view plus any additional event properties or expression values listed as parameters. The synopsis is:

std:size([expression, ...])

The view posts a single long-typed property named size. The view posts the prior size as old data, and the current size as new data to update listeners of the view. Via the GetEnumerator method of the statement the size value can also be polled (read).

As optional parameters the view takes a list of expressions that the view evaluates against the last arriving event and provides along the size field.

An alternative to receiving a data window event count is the prevcount function. Compared to the std:size view the prevcount function requires a data window while the std:size view does not. The related count(...) aggregation function provides a count per group when used with group by.

When combined with a data window view, the size view reports the current number of events in the data window in the insert stream and the prior number of events in the data window as the remove stream. This example reports the number of tick events within the last 1 minute:

select size from StockTickEvent.win:time(1 min).std:size()

To select additional event properties you may add each event property to output as a parameter to the view.

The next example selects the symbol and feed event properties in addition to the size property:

select size, symbol, feed from StockTickEvent.win:time(1 min).std:size(symbol, feed)

The size view is also useful in conjunction with a std:groupwin grouped-window view to count the number of events per group. The EPL below returns the number of events per symbol.

select size from StockTickEvent.std:groupwin(symbol).std:size()

When used without a data window, the view simply counts the number of events:

select size from StockTickEvent.std:size()

All views can be used with pattern statements as well. The next EPL snippet shows a pattern where we look for tick events followed by trade events for the same symbol. The size view counts the number of occurrences of the pattern.

select size from pattern[every s=StockTickEvent -> TradeEvent(symbol=s.symbol)].std:size()

9.2.4. Last Event (std:lastevent)

This view exposes the last element of its parent view:

std:lastevent()

The view acts as a length window of size 1. It thus posts as old events the prior event in the stream, if any.

This example statement retains the last stock tick event for the symbol GE.

select * from StockTickEvent(symbol='GE').std:lastevent()

If you want to output the last event within a sliding window, please see Section 8.1.8, “The Previous Function”. That function accepts a relative (count) or absolute index and returns event properties or an event in the context of the specified data window.

9.2.5. First Event (std:firstevent)

This view retains only the first arriving event:

std:firstevent()

All events arriving after the first event are discarded.

If used within a named window and an on-delete clause deletes the first event, the view resets and will retain the next arriving event.

An example of a statement that retains the first ReferenceData event arriving is:

select * from ReferenceData.std:firstevent()

If you want to output the first event within a sliding window, please see Section 8.1.8, “The Previous Function”. That function accepts a relative (count) or absolute index and returns event properties or an event in the context of the specified data window.

9.2.6. First Unique (std:firstunique)

The firstunique view retains only the very first among events having the same value for the specified expression or list of expressions.

The synopsis is:

std:firstunique(unique_expression [, unique_expression ...])

If used within a named window and an on-delete clause deletes events, the view resets and will retain the next arriving event for the expression result value(s) of the deleted events.

The below example creates a view that retains only the first event per category:

select * from ReferenceData.std:firstunique(category)

9.3. Statistics views

The statistics views can be used combined with data window views or alone. Very similar to aggregation functions, these views aggregate or derive information from an event stream. As compared to aggregation functions, statistics views can post multiple derived fields including properties from the last event that was received. The derived fields and event properties are available for querying in the where-clause and are often compared to prior values using the prior function.

Statistics views accept one or more primary value expressions and any number of optional additional expressions that return values based on the last event received.

9.3.1. Univariate statistics (stat:uni)

This view calculates univariate statistics on a numeric expression. The view takes a single value expression as a parameter plus any number of optional additional expressions to return properties of the last event. The value expression must return a numeric value:

stat:uni(value_expression [,expression, ...])

After the value expression you may optionally list additional expressions or event properties to evaluate for the stream and return their value based on the last arriving event.

Table 9.3. Univariate statistics derived properties

Property NameDescription
datapointsNumber of values, equivalent to count(*) for the stream
totalSum of values
averageAverage of values
varianceVariance
stddevSample standard deviation (square root of variance)
stddevpaPopulation standard deviation

The below example selects the standard deviation on price for stock tick events for the last 10 events.

select stddev from StockTickEvent.win:length(10).stat:uni(price)

To add properties from the event stream you may simply add all additional properties as parameters to the view.

This example selects all of the derived values, based on the price property, plus the values of the symbol and feed event properties:

select * from StockTickEvent.win:length(10).stat:uni(price, symbol, feed)

9.3.2. Regression (stat:linest)

This view calculates regression and related intermediate results on the values returned by two expressions. The view takes two value expressions as parameters plus any number of optional additional expressions to return properties of the last event. The value expressions must return a numeric value:

stat:linest(value_expression, value_expression [,expression, ...])

After the two value expressions you may optionally list additional expressions or event properties to evaluate for the stream and return their value based on the last arriving event.

Table 9.4. Regression derived properties

Property NameDescription
slopeSlope.
YInterceptY intercept.
XAverageX average.
XStandardDeviationPopX standard deviation population.
XStandardDeviationSampleX standard deviation sample.
XSumX sum.
XVarianceX variance.
YAverageX average.
YStandardDeviationPopY standard deviation population.
YStandardDeviationSampleY standard deviation sample.
YSumY sum.
YVarianceY variance.
dataPointsNumber of data points.
nNumber of data points.
sumXSum of X (same as X Sum).
sumXSqSum of X squared.
sumXYSum of X times Y.
sumYSum of Y (same as Y Sum).
sumYSqSum of Y squared.

The next example calculates regression and returns the slope and y-intercept on price and offer for all events in the last 10 seconds.

select slope, YIntercept from StockTickEvent.win:time(10 seconds).stat:linest(price, offer)

To add properties from the event stream you may simply add all additional properties as parameters to the view.

This example selects all of the derived values, based on the price and offer properties, plus the values of the symbol and feed event properties:

select * from StockTickEvent.win:time(10 seconds).stat:linest(price, offer, symbol, feed)

9.3.3. Correlation (stat:correl)

This view calculates the correlation value on the value returned by two expressions. The view takes two value expressions as parameters plus any number of optional additional expressions to return properties of the last event. The value expressions must be return a numeric value:

stat:correl(value_expression, value_expression [,expression, ...])

After the two value expressions you may optionally list additional expressions or event properties to evaluate for the stream and return their value based on the last arriving event.

Table 9.5. Correlation derived properties

Property NameDescription
correlationCorrelation between two event properties

The next example calculates correlation on price and offer over all stock tick events for GE:

select correlation from StockTickEvent(symbol='GE').stat:correl(price, offer)

To add properties from the event stream you may simply add all additional properties as parameters to the view.

This example selects all of the derived values, based on the price and offer property, plus the values of the feed event property:

select * from StockTickEvent(symbol='GE').stat:correl(price, offer, feed)

9.3.4. Weighted average (stat:weighted_avg)

This view returns the weighted average given an expression returning values to compute the average for and an expression returning weight. The view takes two value expressions as parameters plus any number of optional additional expressions to return properties of the last event. The value expressions must return numeric values:

stat:weighted_avg(value_expression_field, value_expression_weight [,expression, ...])

After the value expression you may optionally list additional expressions or event properties to evaluate for the stream and return their value based on the last arriving event.

Table 9.6. Weighted average derived properties

Property NameDescription
averageWeighted average

A statement that derives the volume-weighted average price for the last 3 seconds for a given symbol is shown below:

select average 
from StockTickEvent(symbol='GE').win:time(3 seconds).stat:weighted_avg(price, volume)

To add properties from the event stream you may simply add all additional properties as parameters to the view.

This example selects all of the derived values, based on the price and volume properties, plus the values of the symbol and feed event properties:

select *
from StockTickEvent.win:time(3 seconds).stat:weighted_avg(price, volume, symbol, feed)

Aggregation functions could instead be used to compute the weighted average as well. The next example also posts weighted average per symbol considering the last 3 seconds of stock tick data:

select symbol, sum(price*volume)/sum(volume)
from StockTickEvent.win:time(3 seconds) group by symbol

The following example computes weighted average keeping a separate data window per symbol considering the last 5 events of each symbol:

select symbol, average
from StockTickEvent.std:groupwin(symbol).win:length(5).stat:weighted_avg(price, volume)

9.4. Extension View Set

The views in this set are data windows that order events according to a criteria.

9.4.1. Sorted Window View (ext:sort)

This view sorts by values returned by the specified expressionor list of expressions and keeps only the top (or bottom) events up to the given size.

The syntax is as follows:

ext:sort(size_expression, 
    sort_criteria_expression [asc/desc][, sort_criteria_expression [asc/desc]...]) 

An expression may be followed by the optional asc or desc keywords to indicate that the values returned by that expression are sorted in ascending or descending sort order.

The view below retains only those events that have the highest 10 prices and reports a total price:

select sum(price) from StockTickEvent.ext:sort(10, price desc)

The following example sorts events first by price in descending order, and then by symbol name in ascending (alphabetical) order, keeping only the 10 events with the highest price (with ties resolved by alphabetical order of symbol).

select * from StockTickEvent.ext:sort(10, price desc, symbol asc)

9.4.2. Time-Order View (ext:time_order)

This view orders events that arrive out-of-order, using timestamp-values provided by an expression, and by comparing that timestamp value to engine system time.

The syntax for this view is as follows.

ext:time_order(timestamp_expression, time_period)
ext:time_order(timestamp_expression, seconds_interval_expression)

The first parameter to the view is the expression that supplies timestamp values. The timestamp is expected to be a long-typed millisecond value that denotes an event's time of consideration by the view (or other expression). This is typically the time of arrival. The second parameter is a number-of-seconds expression or the time period specifying the time interval that an arriving event should maximally be held, in order to consider older events arriving at a later time.

Since the view compares timestamp values to engine time, the view requires that the timestamp values and current engine time are both following the same clock. Therefore, to the extend that the clocks that originated both timestamps differ, the view may produce inaccurate results.

As an example, the next statement uses the arrival_time property of MyTimestampedEvent events to order and release events by arrival time:

insert rstream into ArrivalTimeOrderedStream
select rstream * from MyTimestampedEvent.ext:time_order(arrival_time, 10 sec)

In the example above, the arrival_time property holds a long-typed timestamp value in milliseconds. On arrival of an event, the engine compares the timestamp value of each event to the tail-time of the window. The tail-time of the window is, in this example, 10 seconds before engine time (continuously sliding). If the timestamp value indicates that the event is older then the tail-time of the time window, the event is released immediately in the remove stream. If the timestamp value indicates that the event is newer then the tail-time of the window, the view retains the event until engine time moves such that the event timestamp is older then tail-time.

The examples thus holds each arriving event in memory anywhere from zero seconds to 10 seconds, to allow for older events (considering arrival time timestamp) to arrive. In other words, the view holds an event with an arrival time equal to engine time for 10 seconds. The view holds an event with an arrival time that is 2 seconds older then engine time for 8 seconds. The view holds an event with an arrival time that is 10 or more seconds older then engine time for zero seconds, and releases such (old) events immediately into the remove stream.

The insert stream of this sliding window consists of all arriving events. The remove stream of the view is ordered by timestamp value: The event that has the oldest timestamp value is released first, followed by the next newer events. Note the statement above uses the rstream keyword in both the insert into clause and the select clause to select ordered events only. It uses the insert into clause to makes such ordered stream available for subsequent statements to use.

It is up to your application to populate the timestamp property into your events or use a sensible expression that returns timestamp values for consideration by the view. The view also works well if you use externally-provided time via timer events.

Chapter 10. API Reference

10.1. API Overview

Esper has the following primary interfaces:

For EPL introductory information please see Section 4.1, “EPL Introduction” and patterns are described at Section 5.1, “Event Pattern Overview”.

The automated API documentation is also a great source for API information.

10.2. The Service Provider Interface

The EPServiceProvider interface represents an engine instance. Each instance of an Esper engine is completely independent of other engine instances and has its own administrative and runtime interface.

An instance of the Esper engine is obtained via static methods on the EPServiceProviderManager class. The GetDefaultProvider method and the GetProvider(String providerURI) methods return an instance of the Esper engine. The latter can be used to obtain multiple instances of the engine for different provider URI values. The EPServiceProviderManager determines if the provider URI matches all prior provider URI values and returns the same engine instance for the same provider URI value. If the provider URI has not been seen before, it creates a new engine instance.

The code snipped below gets the default instance Esper engine. Subsequent calls to get the default engine instance return the same instance.

EPServiceProvider epService = EPServiceProviderManager.GetDefaultProvider();

This code snippet gets an Esper engine for the provider URI RFIDProcessor1. Subsequent calls to get an engine with the same provider URI return the same instance.

EPServiceProvider epService = EPServiceProviderManager.GetProvider("RFIDProcessor1");

Since the GetProvider methods return the same cached engine instance for each URI, there is no need to statically cache an engine instance in your application.

An existing Esper engine instance can be reset via the Initialize method on the EPServiceProvider instance. This operation stops and removes all statements and resets the engine to the configuration provided when the engine instance for that URI was obtained. If no configuration is provided, an empty (default) configuration applies.

After Initialize your application must obtain new administrative and runtime services. Any administrative and runtime services obtained before the initialize are invalid and have undefined behavior.

The next code snippet outlines a typical sequence of use:

// Configure the engine, this is optional
Configuration config = new Configuration();
config.Configure("configuration.xml");	// load a configuration from file
config.XXX = ...;    // make additional configuration settings

// Obtain an engine instance
EPServiceProvider epService = EPServiceProviderManager.GetDefaultProvider(config);

// Optionally, use initialize if the same engine instance has been used before to start clean
epService.Initialize();

// Optionally, make runtime configuration changes
epService.EPAdministrator.GetConfiguration().Add...(...);

// Destroy the engine instance when no longer needed, frees up resources
epService.Dispose();

An existing Esper engine instance can be destroyed via the Dispose method on the EPServiceProvider instance. This stops and removes all statements as well as frees all resources held by the instance. After a Dispose the engine can no longer be used.

The EPServiceProvider exposes a number of events that are used to inform observers when an engine instance is about to be destroyed, after an engine instance has been initialized, when a new statement gets created and when a statement gets started, stopped or destroyed. Event handlers are registered using the events in the table below.

Table 10.1. Events For Service State Changes

Event on EPServiceProviderDescription 
ServiceDestroyRequested

Occurs after the service has been "destroyed" and all resources have been cleaned up.

 
ServiceInitialized

Occurs after the service provider has been initialized.

 
StatementCreate

Occurs after a statement has been created.

 
StatementStateChange

Occurs when a statement started, stopped or destroyed.

 

As engine instances are completely independent, your application may not send EventBean instances obtained from one engine instance into a second engine instance since the event type space between two engine instances is not shared.

10.3. The Administrative Interface

10.3.1. Creating Statements

Create event pattern expression and EPL statements via the administrative interface EPAdministrator.

This code snippet gets an Esper engine then creates an event pattern and an EPL statement.

EPServiceProvider epService = EPServiceProviderManager.GetDefaultProvider();
EPAdministrator admin = epService.EPAdministrator;

EPStatement 10secRecurTrigger = admin.CreatePattern(
  "every timer:at(*, *, *, *, *, */10)");

EPStatement countStmt = admin.CreateEPL(
  "select count(*) from MarketDataBean.win:time(60 sec)");

Note that event pattern expressions can also occur within EPL statements. This is outlined in more detail in Section 4.4.2, “Pattern-based Event Streams”.

The Create methods on EPAdministrator are overloaded and allow an optional statement name to be passed to the engine. A statement name can be useful for retrieving a statement by name from the engine at a later time. The engine assigns a statement name if no statement name is supplied on statement creation.

The createPattern and CreateEPL methods return EPStatement instances. Statements are automatically started and active when created. A statement can also be stopped and started again via the Stop and Start methods shown in the code snippet below.

countStmt.Stop();
countStmt.Start();

The Create methods on EPAdministrator also accept a user object. The user object is associated with a statement at time of statement creation and is a single, unnamed field that is stored with every statement. Applications may put arbitrary objects in this field. Use the UserObject property on EPStatement to obtain the user object of a statement.

Your application may create new statements or stop and destroy existing statements using any thread and also within event handlers or subscriber code. If using native object events, your application may not create or manage statements in the event object itself while the same event is currently being processed by a statement.

10.3.2. Receiving Statement Results

Esper provides three choices for your application to receive statement results. Your application can use all mechanisms alone or in any combination for each statement. The choices are:

Table 10.2. Choices For Receiving Statement Results

NameMethod, Property or Event on EPStatementDescription
Event HandlersEvents

Your application provides implementations of the UpdateEventHandler to the statement. Event handlers receive UpdateEventArgs instances containing statement results.

The engine continuously indicates results to all event handlers as soon they occur, and following output rate limiting clauses if specified.

Subscriber ObjectSubscriber

Your application provides a vanilla object that exposes methods to receive statement results.

The engine continuously indicates results to the single subscriber as soon they occur, and following output rate limiting clauses if specified.

This is the fastest method to receive statement results, as the engine delivers strongly-typed results directly to your application objects without the need for building an EventBean result set as in the Event Handler choice.

There can be at most 1 Subscriber Object registered per statement. If you require more than one event handler, use the event handler instead (or in addition). The Subscriber Object is bound to the statement with a strongly typed support which ensure direct delivery of new events without type conversion. This optimization is made possible because there can only be 0 or 1 Subscriber Object per statement.

Pull APIGetSafeEnumerator and GetEnumerator

Your application asks the statement for results and receives a set of events via System.Collections.IEnumerable<EventBean>.

This is useful if your application does not need continuous indication of new results in real-time.

Your application may attach one or more event handlers, zero or one single subscriber and in addition use the Pull API on the same statement. There are no limitations to the use of enumerator, subscriber or event handler alone or in combination to receive statement results.

The best delivery performance can generally be achieved by attaching a subscriber and by not attaching event handlers. The engine is aware of the event handlers and subscriber attached to a statement. The engine uses this information internally to reduce statement overhead. For example, if your statement does not have event handlers or a subscriber attached, the engine does not need to continuously generate results for delivery.

If your application attaches both a subscriber and one or more event handlers then the subscriber receives the result first before any of the event handlers.

If your application attaches more then one event handler then the event handlers receive results first in the order they were added to the statement, and To change the order of delivery among event handlers your application can add and remove event handlers at runtime.

If you have configured outbound threading, it means a thread from the outbound thread pool delivers results to the subscriber and event handlers instead of the processing or event-sending thread.

If outbound threading is turned on, we recommend turning off the engine setting preserving the order of events delivered to event handlers as described in Section 11.4.9.1, “Preserving the order of events delivered to listeners”. If outbound threading is turned on statement execution is not blocked for the configured time in the case a subscriber or listener takes too much time.

10.3.3. Setting a Subscriber Object

A subscriber object is a direct binding of query results to a CLR object. The object receives statement results via method invocation. The subscriber class does not need to implement an interface or extend a superclass. Only one subscriber object may be set for a statement.

Subscriber objects have several advantages over event handlers. First, they offer a substantial performance benefit: Query results are delivered directly to your method(s) through method calls, and there is no intermediate representation (EventBean). Second, as subscribers receive strongly-typed parameters, the subscriber code tends to be simpler.

This chapter describes the requirements towards the methods provided by your subscriber class.

The engine can deliver results to your subscriber in two ways:

  1. Each event in the insert stream results in a method invocation, and each event in the remove stream results in further method invocations. This is termed row-by-row delivery.

  2. A single method invocation that delivers all rows of the insert and remove stream. This is termed multi-row delivery.

10.3.3.1. Row-By-Row Delivery

Your subscriber class must provide a method by name Update to receive insert stream events row-by-row. The number and types of parameters declared by the Update method must match the number and types of columns as specified in the select clause, in the same order as in the select clause.

For example, if your statement is:

select orderId, price, count(*) from OrderEvent

Then your subscriber Update method looks as follows:

public class MySubscriber {
  ...
  public void Update(String orderId, double price, long count) {...}
  ...
}

Each method parameter declared by the Update method must be assignable from the respective column type as listed in the select-clause, in the order selected. The assignability rules are:

  • Widening of types follows CLR standards. For example, if your select clause selects an integer value, the method parameter for the same column can be typed int, long, float or double (or any equivalent boxed type).

  • Auto-boxing and unboxing follows CLR standards. For example, if your select clause selects a Nullable<int> value, the method parameter for the same column can be typed int. Note that if your select clause column may generate null values, an exception may occur at runtime unboxing the null value.

  • Interfaces and super-classes are honored in the test for assignability. Therefore System.Object can be used to accept any select clause column type

10.3.3.1.1. Wildcards

If your select clause contains one or more wildcards (*), then the equivalent parameter type is the underlying event type of the stream selected from.

For example, your statement may be:

select *, count(*) from OrderEvent

Then your subscriber Update method looks as follows:

public void Update(OrderEvent orderEvent, long count) {...}

In a join, the wildcard expands to the underlying event type of each stream in the join in the order the streams occur in the from clause. An example statement for a join is:

select *, count(*) from OrderEvent order, OrderHistory hist

Then your subscriber Update method should be:

public void Update(OrderEvent orderEvent, OrderHistory orderHistory, long count) {...}

The stream wildcard syntax and the stream name itself can also be used:

select hist.*, order from OrderEvent order, OrderHistory hist

The matching Update method is:

public void Update(OrderHistory orderHistory, OrderEvent orderEvent) {...}
10.3.3.1.2. Row Delivery as Map and Object Array

Alternatively, your Update method may simply choose to accept System.Collections.Generic.IDictionary as a representation for each row. Each column in the select clause is then made an entry in the resulting Map. The Map keys are the column name if supplied, or the expression string itself for columns without a name.

The Update method for Map delivery is:

public void Update(IDictionary<string, object> row) {...}

The engine also supports delivery of select clause columns as an object array. Each item in the object array represents a column in the select clause. The Update method then looks as follows:

public void Update(Object[] row) {...}
10.3.3.1.3. Delivery of Remove Stream Events

Your subscriber receives remove stream events if it provides a method named UpdateRStream. The method must accept the same number and types of parameters as the Update method.

An example statement:

select orderId, count(*) from OrderEvent.win:time(20 sec) group by orderId

Then your subscriber Update and UpdateRStream methods should be:

public void Update(String, long count) {...}
public void UpdateRStream(String orderId, long count) {...}
10.3.3.1.4. Delivery of Begin and End Indications

If your subscriber requires a notification for begin and end of event delivery, it can expose methods by name UpdateStart and UpdateEnd.

The UpdateStart method must take two integer parameters that indicate the number of events of the insert stream and remove stream to be delivered. The engine invokes the UpdateStart method immediately prior to delivering events to the Update and UpdateRStream methods.

The UpdateEnd method must take no parameters. The engine invokes the UpdateEnd method immediately after delivering events to the Update and UpdateRStream methods.

An example set of delivery methods:

// Called by the engine before delivering events to update methods
public void UpdateStart(int insertStreamLength, int removeStreamLength)

// To deliver insert stream events
public void Update(String orderId, long count) {...}

// To deliver remove stream events
public void UpdateRStream(String orderId, long count) {...}

// Called by the engine after delivering events
public void UpdateEnd() {...}

10.3.3.2. Multi-Row Delivery

In place of row-by-row delivery, your subscriber can receive all events in the insert and remove stream via a single method invocation.

The event delivery follow the scheme as described earlier in Section 10.3.3.1.2, “Row Delivery as Map and Object Array ”. The subscriber class must provide one of the following methods:

Table 10.3. Update Method for Multi-Row Delivery of Underlying Events

MethodDescription
Update(Object[][] insertStream, Object[][] removeStream)

The first dimension of each Object array is the event row, and the second dimension is the column matching the column order of the statement select clause

Update(IDictionary<string,object>[] insertStream, IDictionary<string,object>[] removeStream)

Each map represents one event, and Map entries represent columns of the statement select clause

10.3.3.2.1. Wildcards

If your select clause contains a single wildcard (*) or wildcard stream selector, the subscriber object may also directly receive arrays of the underlying events. In this case, the subscriber class should provide a method Update(Underlying[] insertStream, Underlying[] removeStream) , such that Underlying represents the class of the underlying event.

For example, your statement may be:

select * from OrderEvent.win:time(30 sec)

Your subscriber class exposes the method:

public void Update(OrderEvent[] insertStream, OrderEvent[] removeStream) {...}

10.3.4. Adding Event Handlers

Your application can subscribe to updates posted by a statement via the Events event on EPStatement. Your application must to provide an implementation of the UpdateEventHandler to the statement:

public void HandleEvents(Object sender, UpdateEventArgs updateEventArgs) { ... }
countStmt.Events += HandleEvents;

The following adds and event handler using an embedded lambda expression.

countStmt.Events += (sender, updateEventArgs) => MyMethod(updateEventArgs);

EPL statements and event patterns publish old data and new data to registered UpdateEventHandlers. New data published by statements is the events representing the new values of derived data held by the statement. Old data published by statements constists of the events representing the prior values of derived data held by the statement.

It is important to understand that UpdateEventHandlers receive multiple result rows in one invocation by the engine: the new data and old data parameters to your event handler are array parameters. For example, if your application uses one of the batch data windows, or your application creates a pattern that matches multiple times when a single event arrives, then the engine indicates such multiple result rows in one invocation and your new data array carries two or more rows.

The UpdateEventArgs object that is passed as part of the event handler is especially useful when the same event handler is registered with multiple statements, as the event argumetns contain the statement and engine instance in addition to the new and old data when the engine indicates new results to a event handler.

10.3.4.1. Subscription Snapshot and Atomic Delivery

The AddEventHandlerWithReplay method provided by EPStatement makes it possible to send a snapshot of current statement results to an event handler when the event handler is added.

When using the AddEventHandlerWithReplay method to register an event handler, the event handler receives current statement results as the first invocation of the delegate, passing in the newEvents parameter the current statement results as an array of zero or more events. Subsequent calls to the delegate of the event handler are statement results.

Current statement results are the events returned by the GetEnumerator or GetSafeEnumerator methods.

Delivery is atomic: Events occurring during delivery of current results to the event handler are guaranteed to be delivered in a separate call and not lost. The event handler implementation should thus minimize long-running or blocking operations to reduce lock times held on statement-level resources.

10.3.5. Using Enumerators

Subscribing to events posted by a statement is following a push model. The engine pushes data to event handlers when events are received that cause data to change or patterns to match. Alternatively, you need to know that statements serve up data that your application can obtain via the GetSafeEnumerator and GetEnumerator methods on EPStatement. This is called the pull API and can come in handy if your application is not interested in all new updates, and only needs to perform a frequent or infrequent poll for the latest data.

The GetSafeEnumerator method on EPStatement returns a concurrency-safe enumerator returning current statement results, even while concurrent threads may send events into the engine for processing. The safe enumerator guarantees correct results even as events are being processed by other threads. The cost is that the enumerator obtains and holds a statement lock that must be released via the Dispose method on the GetSafeEnumerator instance.

The GetEnumerator method on EPStatement returns a concurrency-unsafe enumerator. This enumerator is only useful for applications that are single-threaded, or applications that themselves perform coordination between the iterating thread and the threads that send events into the engine for processing. The advantage to this enumerator is that it does not hold a lock.

The next code snippet shows a short example of use of safe iterators:

EPStatement statement = epAdmin.CreateEPL("select avg(price) as avgPrice from MyTick");
// .. send events into the engine
// then use the pull API...
IEnumerator<EventBean> safeEnum = statement.GetSafeEnumerator();
while(safeEnum.MoveNext()) {
 // .. process event ..
 EventBean @event = safeEnum.Current;
 Console.WriteLine("avg:" + @event.Get("avgPrice");
}

This is a short example of use of the regular iterator that is not safe for concurrent event processing:

double averagePrice = (double) eplStatement.FirstOrDefault().Get("average");

The GetSafeEnumerator and GetEnumerator methods can be used to pull results out of all statements, including statements that join streams, contain aggregation functions, pattern statements, and statements that contain a where clause, group by clause, having clause or order by clause.

For statements without an order by clause, the GetEnumerator method returns events in the order maintained by the data window. For statements that contain an order by clause, the GetEnumerator method returns events in the order indicated by the order by clause.

Consider using the on-select clause and a named window if your application requires iterating over a partial result set or requires indexed access for fast iteration; Note that on-select requires that you sent a trigger event, which may contain the key values for indexed access.

Esper places the following restrictions on the pull API and usage of the GetSafeEnumerator and GetEnumerator methods:

  1. In multithreaded applications, use the GetSafeEnumerator method. Note: make sure your application closes the iterator via the Close method when done, otherwise the iterated statement stays locked and event processing for that statement does not resume.

  2. In multithreaded applications, the GetEnumerator method does not hold any locks. The iterator returned by this method does not make any guarantees towards correctness of results and fail-behavior, if your application processes events into the engine instance by multiple threads. Use the GetSafeEnumerator method for concurrency-safe iteration instead.

  3. Since the GetSafeEnumerator and GetEnumerator methods return events to the application immediately, the iterator does not honor an output rate limiting clause, if present. That is, the iterator returns results as if there is no output-rate clause for the statement in statements without grouping or aggregation. For statements with grouping or aggregation, the iterator in combintion with an output clause returns last output group and aggregation results. Use a separate statement and the insert into clause to control the output rate for iteration, if so required.

  4. When iterating a statement that selects an unbound stream (no data window declared), the iterator returns the last event. When iterating a statement that groups and aggregates values from an unbound stream, the iterated result contains only the last updated group.

10.3.6. Managing Statements

The EPAdministrator interface provides the facilities for managing statements:

  • Use the GetStatement method to obtain an existing started or stopped statement by name

  • Use the StatementNames property to obtain a list of started and stopped statement names

  • Use the StartAllStatements, StopAllStatements and DestroyAllStatements methods to manage all statements in one operation

10.3.7. Runtime Configuration

Certain configuration changes are available to perform on an engine instance while in operation. Such configuration operations are available via the GetConfiguration method on EPAdministrator, which returns a ConfigurationOperations object.

Please consult the SDK documentation of ConfigurationOperations for further information. The section Section 11.6, “Runtime Configuration” provides a summary of available configurations.

In summary, the configuration operations available on a running engine instance are as follows:

  • Add new event types for all event representations, check if an event type exists, update an existing Map event type or remove an event type.

  • Add and remove variables (get and set variable values is done via the runtime API).

  • Add a variant stream.

  • Add a revision event type.

  • Add event types for all event classes in a given namespace, using the simple type name as the event name.

  • Add import for user-defined functions.

  • Add a plug-in aggregation function, plug-in single row function, plug-in event type, plug-in event type resolution URIs.

  • Control metrics reporting.

  • Additional items please see the ConfigurationOperations interface.

For examples of above runtime configuration API functions please consider the Configuration chapter, which applies to both static configuration and runtime configuration as the ConfigurationOperations interface is the same.

10.4. The Runtime Interface

The EPRuntime interface is used to send events for processing into an Esper engine, set and get variable values and execute on-demand queries.

The below code snippet shows how to send a CLR object event to the engine. Note that the SendEvent method is overloaded. As events can take on different representation types, the SendEvent takes parameters to reflect the different types of events that can be send into the engine. The Chapter 2, Event Representations section explains the types of events accepted.

EPServiceProvider epService = EPServiceProviderManager.GetDefaultProvider();
EPRuntime runtime = epService.EPRuntime;

// Send an example event containing stock market data
runtime.SendEvent(new MarketDataBean('IBM', 75.0));		

Events, in theoretical terms, are observations of a state change that occurred in the past. Since one cannot change an event that happened in the past, events are best modelled as immutable objects.

Note that the Esper engine relies on events that are sent into an engine to not change their state. Typically, applications create a new event object for every new event, to represent that new event. Application should not modify an existing event that was sent into the engine.

Another important method in the runtime interface is the Route method. This method is designed for use by UpdateEventHandler and subscriber implementations that need to send events into an engine instance to avoid the possibility of a stack overflow due to nested calls to SendEvent.

10.4.1. Event Sender

The EventSender interface processes event objects that are of a known type. This facility can reduce the overhead of event object reflection and type lookup as an event sender is always associated to a single concrete event type.

Use the method GetEventSender(String eventTypeName) to obtain an event sender for processing events of the named type:

EventSender sender = epService.EPRuntime.GetEventSender("MyEvent");
sender.SendEvent(myEvent);

For events backed by a CLR type, the event sender ensures that the event object equals the underlying class, or implements or extends the underlying class for the given event type name.

For events backed by a System.Collections.Generic.IDictionary (Map events), the event sender does not perform any checking other then checking that the event object implements Map.

For events backed by a System.Xml.XmlNode (XML DOM events), the event sender checks that the root element name equals the root element name for the event type.

A second method to obtain an event sender is the method GetEventSender(Uri[]), which takes an array of URIs. This method is for use with plug-in event representations. The event sender returned by this method processes event objects that are of one of the types of one or more plug-in event representations. Please consult Section 13.6, “Custom Event Representation” for more information.

10.4.2. Receiving Unmatched Events

Your application can register an implementation of the UnmatchedListener interface with the EPRuntime runtime via the SetUnmatchedListener method to receive events that were not matched by any statement.

Events that can be unmatched are all events that your application sends into the runtime via one of the SendEvent or Route methods, or that have been generated via an insert into clause.

For an event to become unmatched by any statement, the event must not match any statement's event stream filter criteria. Note that the EPL where clause or having clause are not considered part of the filter criteria for a stream, as explained by example below.

In the next statement all MyEvent events match the statement's event stream filter criteria, regardless of the value of the 'quantity' property. As long as the below statement remains started, the engine would not deliver MyEvent events to your registered UnmatchedListener instance:

select * from MyEvent where quantity > 5

In the following statement a MyEvent event with a 'quantity' property value of 5 or less does not match this statement's event stream filter criteria. The engine delivers such an event to the registered UnmatchedListener instance provided no other statement matches on the event:

select * from MyEvent(quantity > 5)

For patterns, if no pattern sub-expression is active for an event type, an event of that type also counts as unmatched in regards to the pattern statement.

10.4.3. On-Demand Snapshot Query Execution

As your application may not require streaming results and may not know each query in advance, the on-demand query facility provides for ad-hoc execution of an EPL expression.

On-demand queries are not continuous in nature: The query engine executes the query once and returns all result rows to the application. On-demand query execution is very lightweight as the engine performs no statement creation and the query leaves no traces within the engine.

Esper also provides the facility to explicitly index named windows to speed up queries. Please consult Section 4.17.10, “Explicitly Indexing Named Windows” for more information.

The following limitations apply:

  • An on-demand EPL expression only evaluates against the named windows that your application creates. On-demand queries may not specify any other streams or application event types.

  • The following clauses are not allowed in on-demand EPL: insert into and output.

  • Views and patterns are not allowed to appear in on-demand queries.

  • On-demand EPL may not perform subqueries.

  • The previous and prior functions may not be used.

10.4.3.1. On-Demand Query API

The EPRuntime provides two ways to run on-demand queries:

  1. Dynamic on-demand queries are executed once through the ExecuteQuery method.

  2. Prepared on-demand queries: The PrepareQuery method returns an EPOnDemandPreparedQuery representing the query, and the query can be performed repeatedly via the Execute method.

Prepared on-demand queries are designed for repeated execution and may perform better then the dynamic queries if running the same query multiple times. Placeholders are not allowed in prepared on-demand queries.

The next program listing runs an on-demand query against a named window MyNamedWindow and prints a column of each row result of the query:

String query = "select * from MyNamedWindow";
EPOnDemandQueryResult result = epRuntime.executeQuery(query);
for (EventBean row : result.getArray()) {
  System.out.println("name=" + row.get("name"));
}

The next code snippet demonstrates prepared on-demand queries:

EPOnDemandPreparedQuery prepared = epRuntime.prepareQuery(query);
EPOnDemandQueryResult result = prepared.execute();
// ...later ...
prepared.execute();	// execute a second time

Esper also provides the facility to explicitly index named windows to speed up queries. Please consult Section 4.17.10, “Explicitly Indexing Named Windows” for more information.

10.5. Event and Event Type

An EventBean object represents a row (event) in your continuous query's result set. Each EventBean object has an associated EventType object providing event metadata.

An UpdateEventHandler implementation receives one or more UpdateEventArgs events with each invocation. Via the GetEnumerator method on EPStatement your application can poll or read data out of statements. Statement enumerators also return EventBean instances.

Each statement provides the event type of the events it produces, available via the EventType property on EPStatement.

10.5.1. Event Type Metadata

An EventType object encapulates all the metadata about a certain type of events. As Esper supports an inheritance hierarchy for event types, it also provides information about super-types to an event type.

An EventType object provides the following information:

  • For each event property, it lists the property name and type as well as flags for indexed or mapped properties and whether a property is a fragment.

  • The direct and indirect super-types to the event type.

  • Value getters for property expressions.

  • Underlying class of the event representation.

For each property of an event type, there is an EventPropertyDescriptor object that describes the property. The EventPropertyDescriptor contains flags that indicate whether a property is an indexed (array) or a mapped property and whether access to property values require an integer index value (indexed properties only) or string key value (mapped properties only). The descriptor also contains a fragment flag that indicates whether a property value is available as a fragment.

The term fragment means an event property value that is itself an event, or a property value that can be represented as an event. The getFragmentType on EventType may be used to determine a fragment's event type in advance.

A fragment event type and thereby fragment events allow navigation over a statement's results even if the statement result contains nested events or a graph of events. There is no need to use the reflection API to navigate events, since fragments allow the querying of nested event properties or array values, including nested CLR types.

When using the Map event representation, any named Map type nested within a Map as a simple or array property is also available as a fragment. When using CLR objects either directly or within Map events, any object that is neither a primitive or boxed built-in type, and that is not an enumeration and does not implement the Map interface is also available as a fragment.

The nested, indexed and mapped property syntax can be combined to a property expression that may query an event property graph. Most of the methods on the EventType interface allow a property expression to be passed.

Your application may use an EventType object to obtain special getter-objects. A getter-object is a fast accessor to a property value of an event of a given type. All getter objects implement the EventPropertyGetter interface. Getter-objects work only for events of the same type or sub-types as the EventType that provides the EventPropertyGetter. The performance section provides additional information and samples on using getter-objects.

10.5.2. Event Object

An event object is an EventBean that provides:

  • The property value for a property given a property name or property expression that may include nested, indexed or mapped properties in any combination.

  • The event type of the event.

  • Access to the underlying event object.

  • The EventBean fragment or array of EventBean fragments given a property name or property expression.

The GetFragment method on EventBean and EventPropertyGetter return the fragment EventBean or array of EventBean, if the property is itself an event or can be represented as an event. Your application may use EventPropertyDescriptor to determine which properties are also available as fragments.

The underlying event object of an EventBean can be obtained via the Underlying property. Please see Chapter 2, Event Representations for more information on different event representations.

From a threading perspective, it is safe to retain and query EventBean and EventType objects in multiple threads.

10.5.3. Query Example

Consider a statement that returns the symbol, count of events per symbol and average price per symbol for tick events. Our sample statement may declare a fully-qualified type name as the event type: org.sample.StockTickEvent. Assume that this type exists and exposes a symbol property of type String, and a price property of type (primitive) double.

select symbol, avg(price) as avgprice, count(*) as mycount 
from org.sample.StockTickEvent 
group by symbol

The next table summarizes the property names and types as posted by the statement above:

Table 10.4. Properties offered by sample statement aggregating price

NameTypeDescriptionCode snippet
symbolSystem.StringValue of symbol event property
eventBean.Get["symbol"]
avgpriceSystem.DoubleAverage price per symbol
eventBean.Get["avgprice"]
mycountSystem.Int64Number of events per symbol
eventBean["mycount"]

A code snippet out of a possible UpdateEventHandler implementation to this statement may look as below:

String symbol = (String) newEvents[0]["symbol"];
double? price = (double?) newEvents[0]["avgprice"];
long? count= (long?) newEvents[0]["mycount"];

The engine supplies the boxed System.Double? and System.Int64? types as property values rather then primitive types. This is because aggregated values can return a null value to indicate that no data is available for aggregation. Also, in a select statement that computes expressions, the underlying event objects to EventBean instances are of type System.Collections.Generic.IDictionary.

Consider the next statement that specifies a wildcard selecting the same type of event:

select * from org.sample.StockTickEvent where price > 100

The property names and types provided by an EventBean query result row, as posted by the statement above are as follows:

Table 10.5. Properties offered by sample wildcard-select statement

NameTypeDescriptionCode snippet
symbolSystem.StringValue of symbol event property
eventBean["symbol"]
pricedoubleValue of price event property
eventBean["price"]

As an alternative to querying individual event properties via the indexer or Get methods, the Underlying property on EventBean returns the underlying object representing the query result. In the sample statement that features a wildcard-select, the underlying event object is of type org.sample.StockTickEvent:

StockTickEvent tick = (StockTickEvent) newEvents[0].Underlying;

10.5.4. Pattern Example

Composite events are events that aggregate one or more other events. Composite events are typically created by the engine for statements that join two event streams, and for event patterns in which the causal events are retained and reported in a composite event. The example below shows such an event pattern.

// Look for a pattern where BEvent follows AEvent
String pattern = "a=AEvent -> b=BEvent";
EPStatement stmt = epService.EPAdministrator.CreatePattern(pattern);
stmt.Events += HandleEvent;
// Example event handler code
  public void HandleEvent(Object sender, UpdateEventArgs e) {
    Console.WriteLine("a event={0}", e.NewData[0]["a"]);
    Console.WriteLine("b event={0}", e.NewData[0]["b"]);
  }

Note that the HandleEvent method can receive multiple events at once encapsulated in the UpdateEventArgs. For example, a time batch window may post multiple events to event handlers representing a batch of events received during a given time period.

Pattern statements can also produce multiple events delivered to update event handlers in one invocation. The pattern statement below, for instance, delivers an event for each A event that was not followed by a B event with the same id property within 60 seconds of the A event. The engine may deliver all matching A events as an array of events in a single invocation of the UpdateEventHandler delegate associated with the statement:

select * from pattern[
  every a=A -> (timer:interval(60 sec) and not B(id=a.id))]

A code snippet out of a possible UpdateEventHandler implementation to this statement that retrives the events as fragments may look as below:

EventBean a = (EventBean) newEvents[0].GetFragment("a");
// ... or using a nested property expression to get a value out of A event...
double value = (double) newEvent[0]["a.value"];

Some pattern objects return an array of events. An example is the unbound repeat operator. Here is a sample pattern that collects all A events until a B event arrives:

select * from pattern [a=A until b=B]

A possible code to retrieve different fragments or property values:

EventBean[] a = (EventBean[]) newEvents[0].GetFragment("a");
// ... or using a nested property expression to get a value out of A event...
double value = (double) newEvent[0]["a[0].value"];

10.6. Engine Threading and Concurrency

Esper is designed from the ground up to operate as a component to multi-threaded, highly-concurrent applications that require efficient use of resources. In addition, multi-threaded execution requires guarantees in predictability of results and deterministic processing. This section discusses these concerns in detail.

In Esper, an engine instance is a unit of separation. Applications can obtain and discard (initialize) one or more engine instances within the same application domain and can provide the same or different engine configurations to each instance. An engine instance efficiently shares resources between statements. For example, consider two statements that declare the same data window. The engine matches up view declarations provided by each statement and can thus provide a single data window representation shared between the two statements.

Applications can use Esper APIs to concurrently, by multiple threads of execution, perform such functions as creating and managing statements, or sending events into an engine instance for processing. Applications can use application-managed thread pools or any set of same or different threads of execution with any of the public Esper APIs. There are no restrictions towards threading other then those noted in specific sections of this document.

Esper does not prescribe a specific threading model. Applications using Esper retain full control over threading, allowing an engine to be easily embedded and used as a component or library in your favorite container or process.

In the default configuration it is up to the application code to use multiple threads for processing events by the engine, if so desired. All event processing takes places within your application thread call stack. The exception is timer-based processing if your engine instance relies on the internal timer (default). If your application relies on external timer events instead of the internal timer then there need not be any Esper-managed internal threads.

The fact that event processing can take place within your application thread's call stack makes developing applications with Esper easier: Any .NET integrated development environment (IDE) can host an Esper engine instance. This allows developers to easily set up test cases, debug through listener code and inspect input or output events, or trace their call stack.

In the default configuration, each engine instance maintains a single timer thread (internal timer) providing for time or schedule-based processing within the engine. The default resolution at which the internal timer operates is 100 milliseconds. The internal timer thread can be disabled and applications can instead send external time events to an engine instance to perform timer or scheduled processing at the resolution required by an application.

Each engine instance performs minimal locking to enable high levels of concurrency. An engine instance locks on a statement level to protect statement resources.

For an engine instance to produce predictable results from the viewpoint of event handlers to statements, an engine instance by default ensures that it dispatches statement result events to event handlers in the order in which a statement produced result events. Applications that require the highest possible concurrency and do not require predictable order of delivery of events to event handlers, this feature can be turned off via configuration, see Section 11.4.9.1, “Preserving the order of events delivered to listeners”. For example, assume thread T1 processes an event applied to statement S producing output event O1. Assume thread T2 processes another event applied to statement S and produces output event O2. The engine employs a configurable latch system to ensure that event handlers to statement S receive and may complete processing of O1 before receiving O2. When using outbound threading (advanced threading options) or changing the configuration this guarantee is weakened or removed.

In multithreaded environments, when one or more statements make result events available via the insert into clause to further statements, the engine preserves the order of events inserted into the generated insert-into stream, allowing statements that consume other statement's events to behave deterministic. This feature can also be turned off via configuration, see , see Section 11.4.9.2, “Preserving the order of events for insert-into streams”. For example, assume thread T1 processes an event applied to statement S and thread T2 processes another event applied to statement S. Assume statement S inserts into into stream ST. T1 produces an output event O1 for processing by consumers of ST1 and T2 produces an output event O2 for processing by consumers of ST. The engine employs a configurable latch system such that O1 is processed before O2 by consumers of ST. When using route execution threading (advanced threading options) or changing the configuration this guarantee is weakened or removed.

We generally recommended that listener implementations block minimally or do not block at all. By implementing listener code as non-blocking code execution threads can often achieve higher levels of concurrency.

We recommended that, when using a single listener or subscriber instance to receive output from multiple statements, that the listener or subscriber code is multithread-safe. If your application has shared state between listener or subscriber instances then such shared state should be thread-safe.

10.6.1. Advanced Threading

In the default configuration the same application thread that invokes any of the SendEvent methods will process the event fully and also deliver output events to event handlers and subscribers. By default the single internal timer thread based on system time performs time-based processing and delivery of time-based results.

This default configuration reduces the processing overhead associated with thread context switching, is lightweight and fast and works well in many environments such as J2EE, server or client. Latency and throughput requirements are largely use case dependant, and Esper provides engine-level facilities for controlling concurrency that are described next.

Inbound Threading queues all incoming events: A pool of engine-managed threads performs the event processing. The application thread that sends an event via any of the SendEvent methods returns without blocking.

Outbound Threading queues events for delivery to event handlers and subscribers, such that slow or blocking event handlers or subscribers do not block event processing.

Timer Execution Threading means time-based event processing is performed by a pool of engine-managed threads. With this option the internal timer thread (or external timer event) serves only as a metronome, providing units-of-work to the engine-managed threads in the timer execution pool, pushing threading to the level of each statement for time-based execution.

Route Execution Threading means that the thread sending in an event via any of the SendEvent methods (or the inbound threading pooled thread if inbound threading is enabled) only identifies and pre-processes an event, and a pool of engine-managed threads handles the actual processing of the event for each statement, pushing threading to the level of each statement for event-arrival-based execution.

The engine starts engine-managed threads as daemon threads when the engine instance is first obtained. The engine stops engine-managed threads when the engine instance is destroyed via the Dispose method. When the engine is initialized via the Initialize method the existing engine-managed threads are stopped and new threads are created. When shutting down your application, use the Dispose method to stop engine-managed threads.

Note that the options discussed herein may introduce additional processing overhead into your system, as each option involves work queue management and thread context switching.

If your use cases require ordered processing of events or do not tolerate disorder, the threading options described herein may not be the right choice.

If your use cases require loss-less processing of events, wherein the threading options mean that events are held in an in-memory queue, the threading options described herein may not be the right choice.

Care should be taken to consider arrival rates and queue depth. Threading options utilize unbound queues or capacity-bound queues with blocking-put, depending on your configuration, and may therefore introduce an overload or blocking situation to your application. You may use the service provider interface as outlined below to manage queue sizes, if required, and to help tune the engine to your application needs. Consider throttling down the event send rate when the API (see below) indicates that events are getting queued.

All threading options are on the level of an engine. If you require different threading behavior for certain statements then consider using multiple engine instances, consider using the Route method or consider using application threads instead.

Please consult Section 11.4.9, “Engine Settings related to Concurrency and Threading” for instructions on how to configure threading options. Threading options take effect at engine initialization time.

10.6.1.1. Inbound Threading

With inbound threading an engine places inbound events in a queue for processing by one or more engine-managed threads other then the delivering application threads.

The delivering application thread uses one of the SendEvent methods on EPRuntime to deliver events or may also use the SendEvent method on a EventSender. The engine receives the event and places the event into a queue, allowing the delivering thread to continue and not block while the event is being processed and results are delivered.

Events that are sent into the engine via one of the Route methods are not placed into queue but processed by the same thread invoking the Route operation.

10.6.1.2. Outbound Threading

With outbound threading an engine places outbound events in a queue for delivery by one or more engine-managed threads other then the processing thread originating the result.

With outbound threading your listener or subscriber class receives statement results from one of the engine-managed threads in the outbound pool of threads. This is useful when you expect your listener or subscriber code to perform significantly blocking operations and you do not want to hold up event processing.

10.6.1.3. Timer Execution Threading

With timer execution threading an engine places time-based work units into a queue for processing by one or more engine-managed threads other then the internal timer thread or the application thread that sends an external timer event.

Using timer execution threading the internal timer thread (or thread delivering an external timer event) serves to evaluate which time-based work units must be processed. A pool of engine-managed threads performs the actual processing of time-based work units and thereby offloads the work from the internal timer thread (or thread delivering an external timer event).

Enable this option as a tuning parameter when your statements utilize time-based patterns or data windows. Timer execution threading is fine grained and works on the level of a time-based schedule in combination with a statement.

10.6.1.4. Route Execution Threading

With route execution threading an engine identifies event-processing work units based on the event and statement combination. It places such work units into a queue for processing by one or more engine-managed threads other then the thread that originated the event.

While inbound threading works on the level of an event, route execution threading is fine grained and works on the level of an event in combination with a statement.

10.6.1.5. Threading Service Provider Interface

The service-provider interface EPServiceSPI is an extension API that allows to manage engine-level queues and thread pools .

The service-provider interface EPServiceSPI is considered an extension API and subject to change between release versions.

The following code snippet shows how to obtain the BlockingQueue<Runnable> and the ThreadPoolExecutor for the managing the queue and thread pool responsible for inbound threading:

EPServiceProviderSPI spi = (EPServiceProviderSPI) epService;
int queueSize = spi.getThreadingService().getInboundQueue().size();
ThreadPoolExecutor threadpool = spi.getThreadingService().getInboundThreadPool();

10.6.2. Processing Order

10.6.2.1. Competing Statements

This section discusses the order in which N competing statements that all react to the same arriving event execute.

The engine, by default, does not guarantee to execute competing statements in any particular order unless using @Priority. We therefore recommend that an application does not rely on the order of execution of statements by the engine, since that best shields the behavior of an application from changes in the order that statements may get created by your application or by threading configurations that your application may change at will.

If your application requires a defined order of execution of competing statements, use the @Priority EPL syntax to make the order of execution between statements well-defined (requires that you set the prioritized-execution configuration setting). And the @Drop can make a statement preempt all other lowered priority ones that then won't get executed for any matching events.

10.6.2.2. Competing Events in a Work Queue

This section discusses the order of event evaluation when multiple events must be processed, for example when multiple statements use insert-into to generate further events upon arrival of an event.

The engine processes an arriving event completely before considering output events generated by insert-into or routed events inserted by event handlers or subscribers.

For example, assume three statements (1) select * from MyEvent and (2) insert into ABCStream select * from MyEvent. (3) select * from ABCStream. When a MyEvent event arrives then the event handlers to statements (1) and (2) execute first (default threading model). Event handlers to statement (3) which receive the inserted-into stream events are always executed after delivery of the triggering event.

Among all events generated by insert-into of statements and the events routed into the engine via the Route method, all events that insert-into a named window are processed first in the order generated. All other events are processed thereafter in the order they were generated.

When enabling timer or route execution threading as explained under advanced threading options then the engine does not make any guarantee to the processing order except that is will prioritize events inserted into a named window.

10.7. Controlling Time-Keeping

There are two modes for an engine to keep track of time: The internal timer based on JVM system time (the default), and externally-controlled time giving your application full control over the concept of time within an engine or isolated service.

An isolated service is an execution environment separate from the main engine runtime, allowing full control over the concept of time for a group of statements, as further described in Section 10.9, “Service Isolation”.

By default the internal timer provides time and evaluates schedules. External clocking can be used to supply time ticks to the engine instead. The latter is useful for testing time-based event sequences or for synchronizing the engine with an external time source.

The internal timer relies on the windows multimedia timer (default) for time tick events. The windows multimedia timer provides the finest granularity available on a Windows based platform (1 millisecond) but requires integration of the multimedia DLL. The next section describes timer resolution for the internal timer, by default set to 100 milliseconds but is configurable via the threading options. When using externally-controlled time the timer resolution is in your control.

To disable the internal timer and use externally-provided time instead, there are two options. The first option is to use the configuration API at engine initialization time. The second option toggles on and off the internal timer at runtime, via special timer control events that are sent into the engine like any other event.

If using a timer execution thread pool as discussed above, the internal timer or external time event provide the schedule evaluation however do not actually perform the time-based processing. The time-based processing is performed by the threads in the timer execution thread pool.

This code snippet shows the use of the configuration API to disable the internal timer and thereby turn on externally-provided time (see the Configuration section for configuration via XML file):

Configuration config = new Configuration();
config.EngineDefaults.Threading.IsInternalTimerEnabled = false;
EPServiceProvider epService = EPServiceProviderManager.GetDefaultProvider(config);

After disabling the internal timer, it is wise to set a defined time so that any statements created thereafter start relative to the time defined. Use the CurrentTimeEvent class to indicate current time to the engine and to move time forward for the engine.

This code snippet obtains the current time and sends a timer event in:

long timeInMillis = DateTimeHelper.GetCurrentTimeMillis();
CurrentTimeEvent timeEvent = new CurrentTimeEvent(timeInMillis);
epService.EPRuntime.SendEvent(timeEvent);

Alternatively, you can use special timer control events to enable or disable the internal timer. Use the TimerControlEvent class to control timer operation at runtime.

The next code snippet demonstrates toggling to external timer at runtime, by sending in a TimerControlEvent event:

EPServiceProvider epService = EPServiceProviderManager.GetDefaultProvider();
EPRuntime runtime = epService.EPRuntime;
// switch to external clocking
runtime.SendEvent(new TimerControlEvent(TimerControlEvent.ClockType.CLOCK_EXTERNAL));

Your application sends a CurrentTimeEvent event when it desires to move the time forward. All aspects of Esper engine time related to EPL statements and patterns are driven by the time provided by the CurrentTimeEvent that your application sends in.

The next example sequence of instructions sets time to zero, then creates a statement, then moves time forward to 1 seconds later and then 6 seconds later:

// Set start time at zero.
runtime.SendEvent(new CurrentTimeEvent(0));

// create a statement here
epAdministrator.CreateEPL("select * from MyEvent output every 5 seconds");

// move time forward 1 second
runtime.SendEvent(new CurrentTimeEvent(1000));

// move time forward 5 seconds
runtime.SendEvent(new CurrentTimeEvent(6000));

When sending external timer events, your application should make sure that long-type time values are ascending. That is, each long-type value should be either the same value or a larger value then the prior value provided by a CurrentTimeEvent. The engine outputs a warning if time events move back in time.

Your application may use the NextScheduledTime property in EPRuntime to determine the earliest time a schedule for any statement requires evaluation.

The following code snippet sets the current time, creates a statement and prints the next scheduled time which is 1 minute later then the current time:

// Set start time to the current time.
runtime.SendEvent(new CurrentTimeEvent(System.currentTimeMillis()));

// Create a statement.
epService.EPAdministrator.CreateEPL("select * from pattern[timer:interval(1 minute)]");

// Print next schedule time
Console.WriteLine("Next schedule at {0}", new Date(runtime.NextScheduledTime);

10.7.1. Controlling Time Using Time Span Events

With CurrentTimeEvent, as described above, your application can advance engine time to a given point in time. In addition, the NextScheduledTime property in EPRuntime returns the next scheduled time according to started statements. You would typically use CurrentTimeEvent to advance time at a relatively high resolution.

To advance time for a span of time without sending individual CurrentTimeEvent events to the engine, the API provides the class CurrentTimeSpanEvent. You may use CurrentTimeSpanEvent with or without a resolution.

If your application only provides the target end time of time span to CurrentTimeSpanEvent and no resolution, the engine advances time up to the target time by stepping through all relevant times according to started statements.

If your application provides the target end time of time span and in addition a long-typed resolution, the engine advances time up to the target time by incrementing time according to the resolution (regardless of next scheduled time according to started statements).

Consider the following example:

// Set start time to Jan.1, 2010, 00:00 am for this example
DateTime startTime = DateTime.Parse("2010 01 01 00:00:00 000", "yyyy MM dd HH:mm:ss FFF");
runtime.SendEvent(new CurrentTimeEvent(startTime.InMillis()));

// Create a statement.
EPStatement stmt = epService.EPAdministrator.CreateEPL("select current_timestamp() as ct " +
  "from pattern[every timer:interval(1 minute)]");
stmt.Events += ...;	// add an event handler

// Advance time to 10 minutes after start time
runtime.SendEvent(new CurrentTimeSpanEvent(startTime.InMillis() + 10*60*1000));

The above example advances time to 10 minutes after the time set using CurrentTimeSpanEvent. As the example does not pass a resolution, the engine advances time according to statement schedules. Upon sending the CurrentTimeSpanEvent the listener sees 10 invocations for minute 1 to minute 10.

To advance time according to a given resolution, you may provide the resolution as shown below:

// Advance time to 10 minutes after start time at 100 msec resolution
runtime.SendEvent(new CurrentTimeSpanEvent(startTime.getTime() + 10*60*1000, 100));

10.7.2. Additional Time-Related APIs

Consider using the service-provider interface EPRuntimeSPI EPRuntimeIsolatedSPI. The two interfaces are service-provider interfaces that expose additional function to manage statement schedules. However the SPI interfaces should be considered an extension API and are subject to change between release versions.

10.8. Time Resolution

The minimum resolution that all data windows, patterns and output rate limiting operate at is the millisecond. Parameters to time window views, pattern operators or the output clause that are less then 1 millisecond are not allowed. As stated earlier, the default frequency at which the internal timer operates is 100 milliseconds (configurable).

The internal timer thread, by default, uses the call DateTime.Now to obtain system time. Please see the JIRA issue ESPER-191 Support nano/microsecond resolution for more information on system time-call performance, accuracy and drift.

The internal timer thread can be configured to use nano-second time as returned by com.espertech.esper.compat.HighResolutionTimeProvider. If configured for nano-second time, the engine computes an offset of the nano-second ticks to wall clock time upon startup to present back an accurate millisecond wall clock time. Please see section Section 11.4.15, “Engine Settings related to Time Source” to configure the internal timer thread to use HighPerformanceTimer.

Your application can achieve a higher tick rate then 1 tick per millisecond by sending external timer events that carry a long-value which is not based on milliseconds since January 1, 1970, 00:00:00 GMT. In this case, your time interval parameters need to take consideration of the changed use of engine time.

Thus, if your external timer events send long values that represents microseconds (1E-6 sec), then your time window interval must be 1000-times larger, i.e. "win:time(1000)" becomes a 1-second time window.

And therefore, if your external timer events send long values that represents nanoseconds (1E-9 sec), then your time window interval must be 1000000-times larger, i.e. "win:time(1000000)" becomes a 1-second time window.

10.9. Service Isolation

10.9.1. Overview

An isolated service allows an application to control event visibility and the concept of time as desired on a statement level: Events sent into an isolated service are visible only to those statements that currently reside in the isolated service and are not visible to statements outside of that isolated service. Within an isolated service an application can control time independently, start time at a point in time and advance time at the resolution and pace suitable for the statements added to that isolated service.

As discussed before, a single application domain may hold multiple Esper engine instances unique by engine URI. Within an Esper engine instance the default execution environment for statements is the EPRuntime engine runtime, which coordinates all statement's reaction to incoming events and to time passing (via internal or external timer).

Subordinate to an Esper engine instance, your application can additionally allocate multiple isolated services (or execution environments), uniquely identified by a name and represented by the EPServiceProviderIsolated interface. In the isolated service, time passes only when you application sends timer events to the EPRuntimeIsolated instance. Only events explicitly sent to the isolated service are visible to statements added.

Your application can create new statements that start in an isolated service. You can also move existing statements back and forth between the engine and an isolated service.

Isolation does not apply to variables: Variables are global in nature. Also, as named windows are globally visibly data windows, consumers to named windows see changes in named windows even though a consumer or the named window (through the create statement) may be in an isolated service.

An isolated service allows an application to:

  1. Suspend a statement without loosing its statement state that may have accumulated for the statement.

  2. Control the concept of time separately for a set of statements, for example to simulate, backtest, adjust arrival order or compute arrival time.

  3. Initialize statement state by replaying events, without impacting already running statements, to catch-up statements from historical events for example.

While a statement resides in an isolated runtime it receives only those events explicitly sent to the isolated runtime, and performs time-based processing based on the timer events provided to that isolated runtime.

Use the GetEPServiceIsolated method on EPServiceProvider passing a name to obtain an isolated runtime:

EPServiceProviderIsolated isolatedService = epServiceManager.GetEPServiceIsolated("name");

Set the start time for your isolated runtime via the CurrentTimeEvent timer event:

// In this example start the time at the system time
long startInMillis = DateTimeHelper.GetCurrentTimeMillis();	
isolatedService.EPRuntime.SendEvent(new CurrentTimeEvent(startInMillis));

Use the AddStatement method on EPAdministratorIsolated to move an existing statement out of the engine runtime into the isolated runtime:

// look up the existing statement
EPStatement stmt = epServiceManager.EPAdministrator.GetStatement("MyStmt");

// move it to an isolated service
isolatedService.EPAdministrator.AddStatement(stmt);

To remove the statement from isolation and return the statement back to the engine runtime, use the RemoveStatement method on EPAdministratorIsolated:

isolatedService.EPAdministrator.RemoveStatement(stmt);

To create a new statement in the isolated service, use the CreateEPL method on EPAdministratorIsolated:

isolatedService.EPAdministrator.CreateEPL(
  "@Name('MyStmt') select * from Event", null, null); 
// the example is passing the statement name in an annotation and no user object

The Dispose method on EPServiceProviderIsolated moves all currently-isolated statements for that isolated service provider back to engine runtime.

When moving a statement between engine runtime and isolated service or back, the algorithm ensures that events are aged according to the time that passed and time schedules stay intact.

To use isolated services, your configuration must have view sharing disabled as described in Section 11.4.11.1, “Sharing View Resources between Statements”.

10.9.2. Example: Suspending a Statement

By adding an existing statement to an isolated service, the statement's processing effectively becomes suspended. Time does not pass for the statement and it will not process events, unless your application explicitly moves time forward or sends events into the isolated service.

First, let's create a statement and send events:

EPStatement stmt = epServiceManager.EPAdministrator.CreateEPL("select * from TemperatureEvent.win:time(30)");
epServiceManager.EPRuntime.Send(new TemperatureEvent(...));
// send some more events over time

The steps to suspend the previously created statement are as follows:

EPServiceProviderIsolated isolatedService = epServiceManager.GetEPServiceIsolated("suspendedStmts");
isolatedService.EPAdministrator.AddStatement(stmt);

To resume the statement, move the statement back to the engine:

isolatedService.EPAdministrator.RemoveStatement(stmt);

If the statement employed a time window, the events in the time window did not age. If the statement employed patterns, the pattern's time-based schedule remains unchanged. This is because the example did not advance time in the isolated service.

10.9.3. Example: Catching up a Statement from Historical Data

This example creates a statement in the isolated service, replays some events and advances time, then merges back the statement to the engine to let it participate in incoming events and engine time processing.

First, allocate an isolated service and explicitly set it to a start time. Assuming that myStartTime is a long millisecond time value that marks the beginning of the data to replay, the sequence is as follows:

EPServiceProviderIsolated isolatedService = epServiceManager.GetEPServiceIsolated("suspendedStmts");
isolatedService.EPRuntime.SendEvent(new CurrentTimeEvent(myStartTime));

Next, create the statement. The sample statement is a pattern statement looking for temperature events following each other within 60 seconds:

EPStatement stmt = epAdmin.CreateEPL(
  "select * from pattern[every a=TemperatureEvent -> b=TemperatureEvent where timer:within(60)]");

For each historical event to be played, advance time and send an event. This code snippet assumes that currentTime is a time greater then myStartTime and reflects the time that the historical event should be processed at. It also assumes historyEvent is the historical event object.

isolatedService.EPRuntime.SendEvent(new CurrentTimeEvent(currentTime));
isolatedService.EPRuntime.Send(historyEvent);
// repeat the above advancing time until no more events

Finally, when done replaying events, merge the statement back with the engine:

isolatedService.EPAdministrator.RemoveStatement(stmt);

10.9.4. Isolation for Insert-Into

When isolating statements, events that are generated by insert into are visible within the isolated service that currently holds that insert into statement.

For example, assume the below two statements named A and B:

@Name('A') insert into MyStream select * from MyEvent
@Name('B') select * from MyStream

When adding statement A to an isolated service, and assuming a MyEvent is sent to either the engine runtime or the isolated service, a listener to statement B does not receive that event.

When adding statement B to an isolated service, and assuming a MyEvent is sent to either the engine runtime or the isolated service, a listener to statement B does not receive that event.

10.9.5. Isolation for Named Windows

When isolating named windows, the event visibility of events entering and leaving from a named window is not limited to the isolated service. This is because named windows are global data windows (a relation in essence).

For example, assume the below three statements named A, B and C:

@Name('A') create window MyNamedWindow.win:time(60) as select * from MyEvent
@Name('B') insert into MyNamedWindow select * from MyEvent
@Name('C') select * from MyNamedWindow

When adding statement A to an isolated service, and assuming a MyEvent is sent to either the engine runtime or the isolated service, a listener to statement A and C does not receive that event.

When adding statement B to an isolated service, and assuming a MyEvent is sent to either the engine runtime or the isolated service, a listener to statement A and C does not receive that event.

When adding statement C to an isolated service, and assuming a MyEvent is sent to the engine runtime, a listener to statement A and C does receive that event.

10.9.6. Runtime Considerations

Moving statements between an isolated service and the engine is an expensive operation and should not be performed with high frequency.

When using multiple threads to send events and at the same time moving a statement to an isolated service, it its undefined whether events will be delivered to a listener of the isolated statement until all threads completed sending events.

Metrics reporting is not available for statements in an isolated service. Advanced threading options are also not available in the isolated service, however it is thread-safe to send events including timer events from multiple threads to the same or different isolated service.

10.10. Exception Handling

You may register one or more exception handlers for the engine to invoke in the case it encounters an exception processing a continuously-executing statement. By default and without exception handlers the engine cancels execution of the current EPL statement that encountered the exception, logs the exception and continues to the next statement, if any. The configuration is described in Section 11.4.20, “Engine Settings related to Exception Handling”.

If your application registers exception handlers as part of engine configuration, the engine invokes the exception handlers in the order they are registered passing relevant exception information such as EPL statement name, expression and the exception itself.

Exception handlers receive any EPL statement unchecked exception such as internal exceptions or exceptions thrown by plug-in aggregation functions or plug-in views. The engine does not provide to exception handlers any exceptions thrown by static method invocations for function calls, method invocations in joins, methods on event classes and event handlers or subscriber exceptions.

An exception handler can itself throw a runtime exception to cancel execution of the current event against any further statements.

For on-demand queries the API indicates any exception directly back to the caller without the exception handlers being invoked, as exception handlers apply to continuous queries only. The same applies to any API calls other then SendEvent and the EventSender methods.

As the configuration section describes, your application registers one or more classes that implement the ExceptionHandlerFactory interface in the engine configuration. Upon engine initialization the engine obtains a factory instance from the class name that then provides the exception handler instance. The exception handler class must implement the ExceptionHandler interface.

10.11. Condition Handling

You may register one or more condition handlers for the engine to invoke in the case it encounters certain conditions, as outlined below, when executing a statement. By default and without condition handlers the engine logs the condition at informational level and continues processing. The configuration is described in Section 11.4.21, “Engine Settings related to Condition Handling”.

If your application registers condition handlers as part of engine configuration, the engine invokes the condition handlers in the order they are registered passing relevant condition information such as EPL statement name, expression and the condition information itself.

Currently the only condition indicated by this facility is raised by the pattern followed-by operator when used with the limit expression and when the limit is reached, see Section 5.5.8.1, “Limiting Sub-Expression Count”.

A condition handler may not itself throw a runtime exception or return any value.

As the configuration section describes, your application registers one or more classes that implement the ConditionHandlerFactory interface in the engine configuration. Upon engine initialization the engine obtains a factory instance from the class name that then provides the condition handler instance. The condition handler class must implement the ConditionHandler interface.

10.12. Statement Object Model

The statement object model is a set of classes that provide an object-oriented representation of an EPL or pattern statement. The object model classes are found in package com.espertech.esper.client.soda. An instance of EPStatementObjectModel represents a statement's object model.

The statement object model classes are a full and complete specification of a statement. All EPL and pattern constructs including expressions and sub-queries are available via the statement object model.

In conjunction with the administrative API, the statement object model provides the means to build, change or interrogate statements beyond the EPL or pattern syntax string representation. The object graph of the statement object model is fully navigable for easy querying by code, and is also serializable allowing applications to persist or transport statements in object form, when required.

The statement object model supports full round-trip from object model to EPL statement string and back to object model: A statement object model can be rendered into an EPL string representation via the ToEPL method on EPStatementObjectModel. Further, the administrative API allows to compile a statement string into an object model representation via the CompileEPL method on EPAdministrator.

The Create method on EPAdministrator creates and starts a statement as represented by an object model. In order to obtain an object model from an existing statement, obtain the statement expression text of the statement via the Text property on EPStatement and use the CompileEPL method to obtain the object model.

The following limitations apply:

  • Statement object model classes are not safe for sharing between threads other then for read access.

  • Between versions of Esper, the serialized form of the object model is subject to change. Esper makes no guarantees that the serialized object model of one version will be fully compatible with the serialized object model generated by another version of Esper. Please consider this issue when storing Esper object models in persistent store.

10.12.1. Building an Object Model

A EPStatementObjectModel consists of an object graph representing all possible clauses that can be part of an EPL statement.

Among all clauses, the SelectClause and FromClause objects are required clauses that must be present, in order to define what to select and where to select from.

Table 10.6. Required Statement Object Model Instances

ClassDescription
EPStatementObjectModelAll statement clauses for a statement, such as the select-clause and the from-clause, are specified within the object graph of an instance of this class
SelectClauseA list of the selection properties or expressions, or a wildcard
FromClauseA list of one or more streams; A stream can be a filter-based, a pattern-based or a SQL-based stream; Views are added to streams to provide data window or other projections

Part of the statement object model package are convenient builder classes that make it easy to build a new object model or change an existing object model. The SelectClause and FromClause are such builder classes and provide convenient Create methods.

Within the from-clause we have a choice of different streams to select on. The FilterStream class represents a stream that is filled by events of a certain type and that pass an optional filter expression.

We can use the classes introduced above to create a simple statement object model:

EPStatementObjectModel model = new EPStatementObjectModel();
model.SelectClause = SelectClause.CreateWildcard();
model.FromClause = FromClause.Create(FilterStream.Create("com.chipmaker.ReadyEvent"));

The model as above is equivalent to the EPL :

select * from com.chipmaker.ReadyEvent

Last, the code snippet below creates a statement from the object model:

EPStatement stmt = epService.EPAdministrator.Create(model);

Notes on usage:

  • Variable names can simply be treated as property names.

  • When selecting from named windows, the name of the named window is the event type name for use in FilterStream instances or patterns.

  • To compile an arbitrary sub-expression text into an Expression object representation, simply add the expression text to a where clause, compile the EPL string into an object model via the compileEPL on EPAdministrator, and obtain the compiled where from the EPStatementObjectModel via the WhereClause property.

10.12.2. Building Expressions

The EPStatementObjectModel includes an optional where-clause. The where-clause is a filter expression that the engine applies to events in one or more streams. The key interface for all expressions is the Expression interface.

The Expressions class provides a convenient way of obtaining Expression instances for all possible expressions. Please consult the SDK documentatin for detailed method information. The next example discusses sample where-clause expressions.

Use the Expressions class as a service for creating expression instances, and add additional expressions via the Add method that most expressions provide.

In the next example we add a simple where-clause to the EPL as shown earlier:

select * from com.chipmaker.ReadyEvent where line=8

And the code to add a where-clause to the object model is below.

model.WhereClause = Expressions.Eq("line", 8);

The following example considers a more complex where-clause. Assume we need to build an expression using logical-and and logical-or:

select * from com.chipmaker.ReadyEvent 
where (line=8) or (line=10 and age<5)

The code for building such a where-clause by means of the object model classes is:

model.WhereClause = Expressions.Or()
  .Add(Expressions.Eq("line", 8))
  .Add(Expressions.And()
      .Add(Expressions.Eq("line", 10))
      .Add(Expressions.Lt("age", 5))
  );

10.12.3. Building a Pattern Statement

The Patterns class is a factory for building pattern expressions. It provides convenient methods to create all pattern expressions of the pattern language.

Patterns in EPL are seen as a stream of events that consist of patterns matches. The PatternStream class represents a stream of pattern matches and contains a pattern expression within.

For instance, consider the following pattern statement.

select * from pattern [every a=MyAEvent and not b=MyBEvent]

The next code snippet outlines how to use the statement object model and specifically the Patterns class to create a statement object model that is equivalent to the pattern statement above.

EPStatementObjectModel model = new EPStatementObjectModel();
model.SelectClause = SelectClause.CreateWildcard();
PatternExpr pattern = Patterns.And()
  .Add(Patterns.EveryFilter("MyAEvent", "a"))
  .Add(Patterns.NotFilter("MyBEvent", "b"));
model.FromClause = FromClause.Create(PatternStream.Create(pattern));

10.12.4. Building a Select Statement

In this section we build a complete example statement and include all optional clauses in one EPL statement, to demonstrate the object model API.

A sample statement:

insert into ReadyStreamAvg(line, avgAge) 
select line, avg(age) as avgAge 
from com.chipmaker.ReadyEvent(line in (1, 8, 10)).win:time(10) as RE
where RE.waverId != null
group by line 
having avg(age) < 0
output every 10.0 seconds 
order by line

Finally, this code snippet builds the above statement from scratch:

EPStatementObjectModel model = new EPStatementObjectModel();
model.InsertInto = InsertIntoClause.create("ReadyStreamAvg", "line", "avgAge");
model.SelectClause = SelectClause.Create()
    .Add("line")
    .Add(Expressions.Avg("age"), "avgAge");
Filter filter = Filter.Create("com.chipmaker.ReadyEvent", Expressions.In("line", 1, 8, 10));
model.FromClause = FromClause.Create(
    FilterStream.Create(filter, "RE").AddView("win", "time", 10));
model.WhereClause = Expressions.IsNotNull("RE.waverId");
model.GroupByClause = GroupByClause.Create("line");
model.HavingClause = Expressions.Lt(Expressions.Avg("age"), Expressions.Constant(0));
model.OutputLimitClause = OutputLimitClause.Create(OutputLimitSelector.DEFAULT, Expressions.TimePeriod(null, null, null, 10.0, null));
model.OrderByClause = OrderByClause.Create("line");

10.12.5. Building a Create-Variable and On-Set Statement

This sample statement creates a variable:

create variable integer var_output_rate = 10

The code to build the above statement using the object model:

EPStatementObjectModel model = new EPStatementObjectModel();
model.CreateVariable = CreateVariableClause.Create("integer", "var_output_rate", 10));
epService.EPAdministrator.Create(model);

A second statement sets the variable to a new value:

on NewValueEvent set var_output_rate = new_rate

The code to build the above statement using the object model:

EPStatementObjectModel model = new EPStatementObjectModel();
model.OnExpr = OnClause.CreateOnSet("var_output_rate", Expressions.Property("new_rate"));
model.FromClause = FromClause.Create(FilterStream.Create("NewValueEvent"));
EPStatement stmtSet = epService.EPAdministrator.Create(model);

10.12.6. Building Create-Window, On-Delete and On-Select Statements

This sample statement creates a named window:

create window OrdersTimeWindow.win:time(30 sec) as select symbol as sym, volume as vol, price from OrderEvent

The is the code that builds the create-window statement as above:

EPStatementObjectModel model = new EPStatementObjectModel();
model.CreateWindow = CreateWindowClause.Create("OrdersTimeWindow");
model.CreateWindow.AddView("win", "time", 30);
model.SelectClause = SelectClause.Create()
		.AddWithName("symbol", "sym")
		.AddWithName("volume", "vol")
		.Add("price");
model.FromClause = FromClause.Create(FilterStream.Create("OrderEvent));

A second statement deletes from the named window:

on NewOrderEvent as myNewOrders
delete from AllOrdersNamedWindow as myNamedWindow
where myNamedWindow.symbol = myNewOrders.symbol

The object model is built by:

EPStatementObjectModel model = new EPStatementObjectModel();
model.OnExpr = OnClause.CreateOnDelete("AllOrdersNamedWindow", "myNamedWindow");
model.FromClause = FromClause.Create(FilterStream.Create("NewOrderEvent", "myNewOrders"));
model.WhereClause = Expressions.EqProperty("myNamedWindow.symbol", "myNewOrders.symbol");
EPStatement stmtOnDelete = epService.EPAdministrator.Create(model);

A third statement selects from the named window using the non-continuous on-demand selection via on-select:

on QueryEvent(volume>0) as query
select count(*) from OrdersNamedWindow as win
where win.symbol = query.symbol

The on-select statement is built from scratch via the object model as follows:

EPStatementObjectModel model = new EPStatementObjectModel();
model.OnExpr = OnClause.CreateOnSelect("OrdersNamedWindow", "win");
model.WhereClause = Expressions.EqProperty("win.symbol", "query.symbol");
model.FromClause = FromClause.Create(FilterStream.Create("QueryEvent", "query", 
  Expressions.Gt("volume", 0)));
model.SelectClause = SelectClause.Create().Add(Expressions.CountStar());
EPStatement stmtOnSelect = epService.EPAdministrator.Create(model);

10.13. Prepared Statement and Substitution Parameters

The Prepare method that is part of the administrative API pre-compiles an EPL statement and stores the precompiled statement in an EPPreparedStatement object. This object can then be used to efficiently start the parameterized statement multiple times.

Substitution parameters are inserted into an EPL statement as a single question mark character '?'. The engine assigns the first substitution parameter an index of 1 and subsequent parameters increment the index by one.

Substitution parameters can be inserted into any EPL construct that takes an expression. They are therefore valid in any clauses such as the select-clause, from-clause filters, where-clause, group-by-clause, having-clause or order-by-clause, including view parameters and pattern observers and guards. Substitution parameters cannot be used where a numeric constant is required rather then an expression and in SQL statements.

All substitution parameters must be replaced by actual values before a statement with substitution parameters can be started. Substitution parameters can be replaced with an actual value using the SetObject method for each index. Substitution parameters can be set to new values and new statements can be created from the same EPPreparedStatement object more then once.

While the SetObject method allows substitution parameters to assume any actual value including application CLR objects or enumeration values, the application must provide the correct type of substitution parameter that matches the requirements of the expression the parameter resides in.

In the following example of setting parameters on a prepared statement and starting the prepared statement, epService represents an engine instance:

String stmt = "select * from com.chipmaker.ReadyEvent(line=?)";
EPPreparedStatement prepared = epService.EPAdministrator.PrepareEPL(stmt);
prepared.SetObject(1, 8);
EPStatement statement = epService.EPAdministrator.Create(prepared);

10.14. Engine and Statement Metrics Reporting

Metrics reporting is a feature that allows an application to receive ongoing reports about key engine-level and statement-level metrics. Examples are the number of incoming events, the CPU time and wall time taken by statement executions or the number of output events per statement.

Metrics reporting is, by default, disabled. To enable reporting, please follow the steps as outlined in Section 11.4.16, “Engine Settings related to Metrics Reporting”. Metrics reporting must be enabled at engine initialization time. Reporting intervals can be controlled at runtime via the ConfigurationOperations interface available from the administrative API.

Your application receives metrics at configurable intervals via EPL statement. A metric datapoint is simply a well-defined event. The events are EngineMetric and StatementMetric and the type representing the events can be found in the client API in package com.espertech.esper.client.metric.

Since metric events are processed by the engine the same as application events, your EPL may use any construct on such events. For example, your application may select, filter, aggregate properties, sort or insert into a stream or named window all metric events the same as application events.

This example statement selects all engine metric events:

select * from com.espertech.esper.client.metric.EngineMetric

Make sure to have metrics reporting enabled since only then do event handlers or subscribers to a statement such as above receive metric events.

The engine provides metric events after the configured interval of time has passed. By default, only started statements that have activity within an interval (in the form of event or timer processing) are reported upon.

The default configuration performs the publishing of metric events in an Esper daemon thread under the control of the engine instance. Metrics reporting honors externally-supplied time, if using external timer events.

Via runtime configuration options provided by ConfigurationOperations, your application may enable and disable metrics reporting globally, provided that metrics reporting was enabled at initialization time. Your application may also enable and disable metrics reporting for individual statements by statement name.

Statement groups is a configuration feature that allows to assign reporting intervals to statements. Statement groups are described further in the Section 11.4.16, “Engine Settings related to Metrics Reporting” section. Statement groups cannot be added or removed at runtime.

The following limitations apply:

  • High-performance numbers are obtained through the use of a PerformanceObserver that uses the Kernel32 DLL. While this method has been exhaustively tested on Windows, we are aware that this will not work in non-Windows environments. To ensure compatability with Mono, we have included a model based on DateTime. This implementation will not provide the level of granularity or clarity that the Windows implementation will.

  • There is a performance cost to collecting and reporting metrics.

  • Not all statements may report metrics: The engine performs certain runtime optimizations sharing resources between similar statements, thereby not reporting on certain statements unless resource sharing is disabled through configuration.

10.14.1. Engine Metrics

Engine metrics are properties of EngineMetric events:

Table 10.7. EngineMetric Properties

NameDescription
engineURIThe URI of the engine instance.
timestampThe current engine time.
inputCountCumulative number of input events since engine initialization time. Input events are defined as events send in via application threads as well as insert into events.
scheduleDepthNumber of outstanding schedules.

10.14.2. Statement Metrics

Statement metrics are properties of StatementMetric. The properties are:

Table 10.8. StatementMetric Properties

NameDescription
engineURIThe URI of the engine instance.
timestampThe current engine time.
statementNameStatement name, if provided at time of statement creation, otherwise a generated name.
cpuTimeStatement processing CPU time (system and user) in nanoseconds.
wallTimeStatement processing wall time in nanoseconds (based on System.nanoTime).
numOutputIStreamNumber of insert stream rows output to event handlers or the subscriber, if any.
numOutputRStreamNumber of remove stream rows output to event handlers or the subscriber, if any.

The totals reported are cumulative relative to the last metric report.

10.15. Event Rendering to XML and JSON

Your application may use the built-in XML and JSON formatters to render output events into a readable textual format, such as for integration or debugging purposes. This section introduces the utility classes in the client util package for rendering events to strings. Further API information can be found in the SDK documentation.

The IEventRenderer interface accessible from the runtime interface via the GetEventRenderer method provides access to JSON and XML rendering. For repeated rendering of events of the same event type or subtypes, it is recommended to obtain a JSONEventRenderer or XMLEventRenderer instance and use the Render method provided by the interface. This allows the renderer implementations to cache event type metadata for fast rendering.

In this example we show how one may obtain a renderer for repeated rendering of events of the same type, assuming that statement is an instance of EPStatement:

JSONEventRenderer jsonRenderer = epService.EPRuntime.
    EventRenderer.GetJSONRenderer(statement.EventType);

Assuming that event is an instance of EventBean, this code snippet renders an event into the JSON format:

String jsonEventText = jsonRenderer.Render("MyEvent", @event);

The XML renderer works the same:

XMLEventRenderer xmlRenderer = epService.EPRuntime.
    EventRenderer.GetXMLRenderer(statement.EventType);

...and...

String xmlEventText = xmlRenderer.Render("MyEvent", @event);

If the event type is not known in advance or if you application does not want to obtain a renderer instance per event type for fast rendering, your application can use one of the following methods to render an event to a XML or JSON textual format:

String json = epService.EPRuntime.EventRenderer.RenderJSON(@event);
String xml = epService.EPRuntime.EventRenderer.RenderXML(@event);

Use the JSONRenderingOptions or XMLRenderingOptions classes to control how events are rendered.

10.15.1. JSON Event Rendering Conventions and Options

The JSON renderer produces JSON text according to the standard documented at http://www.json.org.

The renderer formats simple properties as well as nested properties and indexed properties according to the JSON string encoding, array encoding and nested object encoding requirements.

The renderer does render indexed properties, it does not render indexed properties that require an index, i.e. if your event representation is backed by native objects and your getter method is getValue(int index), the indexed property values are not part of the JSON text. This is because the implementation has no way to determine how many index keys there are. A workaround is to have a method such as Object[] getValue() instead.

The same is true for mapped properties that the renderer also renders. If a property requires a Map key for access, i.e. your getter method is getValue(String key), such property values are not part of the result text as there is no way for the implementation to determine the key set.

10.15.2. XML Event Rendering Conventions and Options

The XML renderer produces well-formed XML text according to the XML standard.

The renderer can be configured to format simple properties as attributes or as elements. Nested properties and indexed properties are always represented as XML sub-elements to the root or parent element.

The root element name provided to the XML renderer must be the element name of the root in the XML document and may include namespace instructions.

The renderer does render indexed properties, it does not render indexed properties that require an index, i.e. if your event representation is backed by native objects and your getter method is GetValue(int index), the indexed property values are not part of the XML text. This is because the implementation has no way to determine how many index keys there are. A workaround is to have a method such as Object[] GetValue() instead.

The same is true for mapped properties that the renderer also renders. If a property requires a Map key for access, i.e. your getter method is GetValue(String key), such property values are not part of the result text as there is no way for the implementation to determine the key set.

10.16. Plug-in Loader

A plug-in loader is for general use with input adapters, output adapters or EPL code deployment or any other task that can benefits from being part of an Esper configuration file and that follows engine lifecycle.

A plug-in loader implements the com.espertech.esper.plugin.PluginLoader interface and can be listed in the configuration.

Each configured plug-in loader follows the engine instance lifecycle: When an engine instance initializes, it instantiates each PluginLoader implementation class listed in the configuration. The engine then invokes the lifecycle methods of the PluginLoader implementation class before and after the engine is fully initialized and before an engine instance is destroyed.

Declare a plug-in loader in your configuration XML as follows:

...
  <plugin-loader name="MyLoader" class-name="org.mypackage.MyLoader">
    <init-arg name="property1" value="val1"/>
  </plugin-loader>
...

Alternatively, add the plug-in loader via the configuration API:

Configuration config = new Configuration();
var props = new Dictionary<string, string>();
props["property1"] = "value1";
config.AddPluginLoader("MyLoader", "org.mypackage.MyLoader", props);

Implement the Init method of your PluginLoader implementation to receive initialization parameters. The engine invokes this method before the engine is fully initialized, therefore your implementation should not yet rely on the engine instance within the method body:

public class MyPluginLoader : PluginLoader {
  public void Init(String loaderName, Properties properties, EPServiceProviderSPI epService) {
     // save the configuration for later, perform checking
  }
  ...

The engine calls the PostInitialize method once the engine completed initialization and to indicate the engine is ready for traffic.

public void PostInitialize() {
  // Start the actual interaction with external feeds or the engine here
}
...

The engine calls the Dispose method once the engine is destroyed or initialized for a second time.

public void Dispose() {
  // Destroy resources allocated as the engine instance is being destroyed
}

10.17. CLR-Java Differences

When using Esper under .NET, one is sure to run into some differences that make it noticably different that Java. For example, Java does not have an explicit notion of properties; however, this concept is native to the CLR. Another difference is in how events are handled; in Java, it is common to expose an interface that is implemented by the event receiver; in the CLR, its common to expose an event handler that requires a delegate that the event receiver provides.

Chapter 11. Configuration

Esper engine configuration is entirely optional. Esper has a very small number of configuration parameters that can be used to simplify event pattern and EPL statements, and to tune the engine behavior to specific requirements. The Esper engine works out-of-the-box without configuration.

An application can supply configuration at the time of engine allocation using the Configuration class, and can also use XML files to hold configuration. Configuration can be changed at runtime via the ConfigurationOperations interface available from EPAdministrator via the GetConfiguration method.

The difference between using a Configuration object and the ConfigurationOperations interface is that for the latter, all configuration including event types added through that interface are considered runtime configurations. This means they will be discarded when calling the Initialize method on an EPServiceProvider instance.

11.1. Programmatic Configuration

An instance of com.espertech.esper.client.Configuration represents all configuration parameters. The Configuration is used to build an EPServiceProvider, which provides the administrative and runtime interfaces for an Esper engine instance.

You may obtain a Configuration instance by instantiating it directly and adding or setting values on it. The Configuration instance is then passed to EPServiceProviderManager to obtain a configured Esper engine.

Configuration configuration = new Configuration();
configuration.AddEventType<PriceLimit>("PriceLimit");
configuration.AddEventType<StockTick>("StockTick");
configuration.AddImport("org.mycompany.mypackage.MyUtility");
configuration.AddImport("org.mycompany.util");

EPServiceProvider epService = EPServiceProviderManager.GetProvider("sample", configuration);

Note that Configuration is meant only as an initialization-time object. The Esper engine represented by an EPServiceProvider does not retain any association back to the Configuration.

The ConfigurationOperations interface provides runtime configuration options as further described in Section 10.3.7, “Runtime Configuration”. Through this interface applications can, for example, add new event types at runtime and then create new statements that rely on the additional configuration. The GetConfiguration method on EPAdministrator allows access to ConfigurationOperations.

11.2. Configuration via XML File

An alternative approach to configuration is to specify a configuration in a XML file.

The default name for the XML configuration file is esper.cfg.xml. Esper reads this file from the root of the CLASSPATH as an application resource via the Configure method.

Configuration configuration = new Configuration();		
configuration.Configure();

The Configuration class can read the XML configuration file from other sources as well. The Configure method accepts URL, File and String filename parameters.

Configuration configuration = new Configuration();		
configuration.Configure("myengine.esper.cfg.xml");

11.3. XML Configuration File

Here is an example configuration file. The schema for the configuration file can be found in the etc folder and is named esper-configuration-4-0.xsd. It is also available online at http://www.espertech.com/schema/esper/esper-configuration-2.0.xsd so that IDE can fetch it automatically. The namespace used is http://www.espertech.com/schema/esper.

<?xml version="1.0" encoding="UTF-8"?>
<esper-configuration xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xmlns="http://www.espertech.com/schema/esper"
    xsi:schemaLocation="
http://www.espertech.com/schema/esper
http://www.espertech.com/schema/esper/esper-configuration-2.0.xsd">
  <event-type name="StockTick" class="com.espertech.esper.example.stockticker.event.StockTick"/>
  <event-type name="PriceLimit" class="com.espertech.esper.example.stockticker.event.PriceLimit"/>
  <auto-import import-name="org.mycompany.mypackage.MyUtility"/>
  <auto-import import-name="org.mycompany.util.*"/>
</esper-configuration>		

The example above is only a subset of the configuration items available. The next chapters outline the available configuration in greater detail.

11.4. Configuration Items

11.4.1. Events represented by CLR types

11.4.1.1. Package of CLR Event Classes

Via this configuration an application can make the Java package or packages that contain an application's Java event classes known to an engine. Thereby an application can simply refer to event types in statements by using the simple class name of each CLR type representing an event type.

For example, consider an order-taking application that places all event classes in package com.mycompany.order.event. One CLR type representing an event is the class OrderEvent. The application can simply issue a statement as follows to select OrderEvent events:

select * from OrderEvent

The XML configuration for defining the namespaces that contain CLR event types is:

<event-type-auto-name package-name="com.mycompany.order.event"/>

The same configuration but using the Configuration class:

Configuration config = new Configuration();
config.AddEventTypeAutoName("com.mycompany.order.events");
// ... or ...
config.AddEventTypeAutoName(typeof(MyEvent).Namespace);

11.4.1.2. Event type name to CLR type mapping

This configuration item can be used to allow event pattern statements and EPL statements to use an event type name rather then the fully qualified CLR type name. Note that interface classes and abstract classes are also supported as event types via the fully qualified CLR type name, and an event type name can also be defined for such classes.

The example pattern statement below first shows a pattern that uses the name StockTick. The second pattern statement is equivalent but specifies the fully-qualified CLR type name.

every StockTick(symbol='IBM')"
every com.espertech.esper.example.stockticker.event.StockTick(symbol='IBM')

The event type name can be listed in the XML configuration file as shown below. The Configuration API can also be used to programatically specify an event type name, as shown in an earlier code snippet.

<event-type name="StockTick" class="com.espertech.esper.example.stockticker.event.StockTick"/>

11.4.1.3. Legacy Event Classes

Esper can process CLR typees that provide event properties through other means then through JavaBean-style getter methods. It is not necessary that the method and member variable names in your CLR type adhere to the JavaBean convention - any public methods and public member variables can be exposed as event properties via the below configuration.

A CLR type can optionally be configured with an accessor style attribute. This attribute instructs the engine how it should expose methods and fields for use as event properties in statements.

Table 11.1. Accessor Styles

Style NameDescription
nativeAs the default setting, the engine exposes an event property for each public readable property and for all methods that expose properties according to the simple property publisher specification.
publicThe engine exposes an event property for each public method and public member variable of the given class
explicitThe engine exposes an event property only for the explicitly configured public methods and public member variables

Using the public setting for the accessor-style attribute instructs the engine to expose an event property for each public method and public member variable of a CLR type. The engine assigns event property names of the same name as the name of the method or member variable in the CLR type.

For example, assuming the class MyLegacyEvent exposes a method named readValue and a member variable named myField, we can then use properties as shown.

select readValue, myField from MyLegacyEvent

Using the explicit setting for the accessor-style attribute requires that event properties are declared via configuration. This is outlined in the next chapter.

When configuring an engine instance from a XML configuration file, the XML snippet below demonstrates the use of the legacy-type element and the accessor-style attribute.

<event-type name="MyLegacyEvent" class="com.mycompany.mypackage.MyLegacyEventClass">
  <legacy-type accessor-style="public"/>
</event-type>

When configuring an engine instance via Configuration API, the sample code below shows how to set the accessor style.

Configuration configuration = new Configuration();
ConfigurationEventTypeLegacy legacyDef = new ConfigurationEventTypeLegacy();
legacyDef.setAccessorStyle(ConfigurationEventTypeLegacy.AccessorStyle.PUBLIC);
config.AddEventType("MyLegacyEvent", MyLegacyEventClass.class.getName(), legacyDef);

EPServiceProvider epService = EPServiceProviderManager.getProvider("sample", configuration);

11.4.1.4. Specifying Event Properties for CLR Types

Sometimes it may be convenient to use event property names in pattern and EPL statements that are backed up by a given public method or member variable (field) in a CLR type. And it can be useful to declare multiple event properties that each map to the same method or member variable.

We can configure properties of events via method-property and field-property elements, as the next example shows.

<event-type name="StockTick" class="com.espertech.esper.example.stockticker.event.StockTickEvent">
	<legacy-type accessor-style="native" code-generation="enabled">
		<method-property name="price" accessor-method="getCurrentPrice" />
		<field-property name="volume" accessor-field="volumeField" />
	</legacy-type>
</event-type>

The XML configuration snippet above declared an event property named price backed by a getter-method named getCurrentPrice, and a second event property named volume that is backed by a public member variable named volumeField. Thus the price and volume properties can be used in a statement:

select avg(price * volume) from StockTick

As with all configuration options, the API can also be used:

Configuration configuration = new Configuration();
ConfigurationEventTypeLegacy legacyDef = new ConfigurationEventTypeLegacy();
legacyDef.AddMethodProperty("price", "GetCurrentPrice");
legacyDef.AddFieldProperty("volume", "_volumeField");
config.AddEventType<StockTickEvent>("StockTick", legacyDef);

11.4.1.5. Turning off Code Generation

Esper employes the CGLIB library for very fast read access to event property values. For certain legacy CLR typees it may be desirable to disable the use of this library and instead use CLR reflection to obtain event property values from event objects.

In the XML configuration, the optional code-generation attribute in the legacy-type section can be set to disabled as shown next.

<event-type name="MyLegacyEvent" class="com.mycompany.package.MyLegacyEventClass">
	<legacy-type accessor-style="native" code-generation="disabled" />
</event-type>

The sample below shows how to configure this option via the API.

Configuration configuration = new Configuration();
ConfigurationEventTypeLegacy legacyDef = new ConfigurationEventTypeLegacy();
legacyDef.CodeGeneration = ConfigurationEventTypeLegacy.CodeGeneration.DISABLED;
config.AddEventType<MyLegacyEventClass>("MyLegacyEvent", legacyDef);

11.4.1.6. Case Sensitivity and Property Names

By default the engine resolves Java event properties case sensitive. That is, property names in statements must match property names in name and case. This option controls case sensitivity per CLR type.

In the configuration XML, the optional property-resolution-style attribute in the legacy-type element can be set to any of these values:

Table 11.2. Property Resolution Case Sensitivity Styles

Style NameDescription
case_sensitive (default)As the default setting, the engine matches property names for the exact name and case only.
case_insensitiveProperties are matched if the names are identical. A case insensitive search is used and will choose the first property that matches the name exactly or the first property that matches case insensitively should no match be found.
distinct_case_insensitiveProperties are matched if the names are identical. A case insensitive search is used and will choose the first property that matches the name exactly case insensitively. If more than one 'name' can be mapped to the property an exception is thrown.

The sample below shows this option in XML configuration, however the setting can also be changed via API:

<event-type name="MyLegacyEvent" class="com.mycompany.package.MyLegacyEventClass">
  <legacy-type property-resolution-style="case_insensitive"/>
</event-type>

11.4.1.7. Factory and Copy Method

The insert into clause and directly instantiate and populate your event object. By default the engine invokes the default constructor to instantiate an event object. To change this behavior, you may configure a factory method. The factory method is a method name or a class name plus a method name (in the format class.method) that returns an instance of the class.

The update clause can change event properties on an event object. For the purpose of maintaining consistency, the engine may have to copy your event object via serialization (annotated with the System.Serializable attribute). If instead you do not want any copy operations to occur, or your application needs to control the copy operation, you may configure a copy method. The copy method is the name of a method on the event object that copies the event object.

The sample below shows this option in XML configuration, however the setting can also be changed via API:

<event-type name="MyLegacyEvent" class="com.mycompany.package.MyLegacyEventClass"
  factory-method="com.mycompany.myapp.MySampleEventFactory.createMyLegacyTypeEvent" copy-method="myCopyMethod"> 
</event-type>

11.4.2. Events represented by System.Collection.Generic.IDictionary

The engine can process System.Collection.Generic.IDictionary events via the SendEvent(Map map, String eventTypeName) method on the EPRuntime interface. Entries in the Map represent event properties. Keys must be of type System.String for the engine to be able to look up event property names in pattern or EPL statements. Values can be of any type. CLR objects as values in a Map can be processed by the engine, and strongly-typed nested maps are also supported. Please see the Chapter 2, Event Representations section for details on how to use Map events with the engine.

Via configuration we provide an event type name for Map events for use in statements, and the event property names and types enabling the engine to validate properties in statements.

The below snippet of XML configuration configures an event named MyMapEvent.

<event-type name="MyMapEvent">
  <clr-map>
    <map-property name="carId" class="int"/>
    <map-property name="carType" class="string"/>
    <map-property name="assembly" class="com.mycompany.Assembly"/>    
  </clr-map>
</event-type>

This configuration defines the carId property of MyMapEvent events to be of type int, and the carType property to be of type System.String. The assembly property of the Map event will contain instances of com.mycompany.Assembly for the engine to query.

The valid types for the class attribute are listed in Section 11.5, “Type Names”. In addition, any fully-qualified CLR type name that can be resolved via Class.forName is allowed.

You can also use the configuration API to configure Map event types, as the short code snippet below demonstrates:

IDictionary<String, Object> properties = new Dictionary<String, Object>();
properties["carId"] = "int";
properties["carType"] = "string";
properties["assembly"] = typeof(Assembly).FullName;

Configuration configuration = new Configuration();
configuration.AddEventType("MyMapEvent", properties);

For strongly-typed nested maps (maps-within-maps), the configuration API method addEventType can also used to define the nested types. The XML configuration does not provide the capability to configure nested maps.

Finally, here is a sample EPL statement that uses the configured MyMapEvent map event. This statement uses the chassisTag and numParts properties of Assembly objects in each map.

select carType, assembly.chassisTag, count(assembly.numParts) from MyMapEvent.win:time(60 sec)

A Map event type may also become a subtype of one or more supertypes that must also be Map event types. The clr-map element provides an optional attribute supertype-names that accepts a comma-separated list of names of Map event types that are supertypes to the type:

<event-type name="AccountUpdate">
<clr-map supertype-names="BaseUpdate, AccountEvent">
...

For initialization time configuration, the AddMapSuperType method can be used to add Map hierarchy information. For runtime configuration, pass the supertype names to the AddEventType method in ConfigurationOperations.

11.4.3. Events represented by System.Xml.XmlNode

Via this configuration item the Esper engine can natively process System.Xml.XmlNode instances, i.e. XML document object model (DOM) nodes. Please see the Chapter 2, Event Representations section for details on how to use Node events with the engine.

Esper allows configuring XPath expressions as event properties. You can specify arbitrary XPath functions or expressions and provide a property name by which their result values will be available for use in expressions.

For XML documents that follow a XML schema, Esper can load and interrogate your schema and validate event property names and types against the schema information.

Nested, mapped and indexed event properties are also supported in expressions against System.Xml.XmlNode events. Thus XML trees can conveniently be interrogated using the existing event property syntax for querying JavaBean objects, JavaBean object graphs or System.Collection.Generic.IDictionary events.

In the simplest form, the Esper engine only requires a configuration entry containing the root element name and the event type name in order to process System.Xml.XmlNode events:

<event-type name="MyXMLNodeEvent">
  <xml-dom root-element-name="myevent" />
</event-type>

You can also use the configuration API to configure XML event types, as the short example below demonstrates. In fact, all configuration options available through XML configuration can also be provided via setter methods on the ConfigurationEventTypeXMLDOM class.

Configuration configuration = new Configuration();
ConfigurationEventTypeXMLDOM desc = new ConfigurationEventTypeXMLDOM();
desc.RootElementName = "myevent";
desc.AddXPathProperty("name1", "/element/@attribute", XPathResultType.String);
desc.AddXPathProperty("name2", "/element/subelement", XPathResultType.Number);
configuration.addEventType("MyXMLNodeEvent", desc);

The next example presents configuration options in a sample configuration entry.

<event-type name="AutoIdRFIDEvent">
  <xml-dom root-element-name="Sensor" schema-resource="data/AutoIdPmlCore.xsd" 
       default-namespace="urn:autoid:specification:interchange:PMLCore:xml:schema:1">
    <namespace-prefix prefix="pmlcore" 
       namespace="urn:autoid:specification:interchange:PMLCore:xml:schema:1"/>
    <xpath-property property-name="countTags" 
       xpath="count(/pmlcore:Sensor/pmlcore:Observation/pmlcore:Tag)" type="number"/>
  </xml-dom>
</event-type>

This example configures an event property named countTags whose value is computed by an XPath expression. The namespace prefixes and default namespace are for use with XPath expressions and must also be made known to the engine in order for the engine to compile XPath expressions. Via the schema-resource attribute we instruct the engine to load a schema file. You may also use schema-text instead to provide the actual text of the schema.

Here is an example EPL statement using the configured event type named AutoIdRFIDEvent.

select ID, countTags from AutoIdRFIDEvent.win:time(30 sec)

11.4.3.1. Schema Resource

The schema-resource attribute takes a schema resource URL or classpath-relative filename. The engine attempts to resolve the schema resource as an URL. If the schema resource name is not a valid URL, the engine attempts to resolve the resource from resource manager (com.espertech.esper.compat.ResourceManager). Use the schema-text attribute instead when it is more practical to provide the actual text of the schema.

By configuring a schema file for the engine to load, the engine performs these additional services:

  • Validates the event properties in a statement, ensuring the event property name matches an attribute or element in the XML

  • Determines the type of the event property allowing event properties to be used in type-sensitive expressions such as expressions involving arithmetic (Note: XPath properties are also typed)

  • Matches event property names to either element names or attributes

If no schema resource is specified, none of the event properties specified in statements are validated at statement creation time and their type defaults to System.String. Also, attributes are not supported if no schema resource is specified and must thus be declared via XPath expression.

11.4.3.2. Explicit XPath Property

The xpath-property element adds explicitly-names event properties to the event type that are computed via an XPath expression. In order for the XPath expression to compile, be sure to specify the default-namespace attribute and use the namespace-prefix to declare namespace prefixes.

XPath expression properties are strongly typed. The type attribute allows the following values. These values correspond to those declared by System.Xml.XPath.XPathResultType.

  • number (Note: resolves to a double)

  • string

  • boolean

  • any

  • nodeset

In case you need your XPath expression to return a type other then the types listed above, an optional cast-to type can be specified. If specified, the operation firsts obtains the result of the XPath expression as the defined type (number, string, boolean) and then casts or parses the returned type to the specified cast-to-type. At runtime, a warning message is logged if the XPath expression returns a result object that cannot be casted or parsed.

The next line shows how to return a long-type property for an XPath expression that returns a string:

desc.AddXPathProperty("name", "/element/sub", XPathConstants.STRING, "long");

The equivalent configuration XML is:

<xpath-property property-name="name"  xpath="/element/sub" type="string" cast="long"/>

See Section 11.5, “Type Names” for a list of cast-to type names.

11.4.3.3. Absolute or Deep Property Resolution

This setting indicates that when properties are compiled to XPath expressions that the compilation should generate an absolute XPath expression or a deep (find element) XPath expression.

For example, consider the following statement against an event type that is represented by a XML DOM document, assuming the event type GetQuote has been configured with the engine as a XML DOM event type:

select request, request.symbol from GetQuote

By default, the engine compiles the "request" property name to an XPath expression "/GetQuote/request". It compiles the nested property named "request.symbol" to an XPath expression "/GetQuote/request/symbol", wherein the root element node is "GetQuote".

By setting absolute property resolution to false, the engine compiles the "request" property name to an XPath expression "//request". It compiles the nested property named "request.symbol" to an XPath expression "//request/symbol". This enables these elements to be located anywhere in the XML document.

The setting is available in XML via the attribute resolve-properties-absolute.

The configuration API provides the above settings as shown here in a sample code:

ConfigurationEventTypeXMLDOM desc = new ConfigurationEventTypeXMLDOM();
desc.RootElementName = "GetQuote";
desc.DefaultNamespace = "http://services.samples/xsd";
desc.RootElementNamespace = "http://services.samples/xsd";
desc.AddNamespacePrefix("m0", "http://services.samples/xsd");
desc.IsXPathResolvePropertiesAbsolute = false;
configuration.AddEventType("GetQuote", desc);

11.4.3.4. XPath Variable and Function Resolver

If your XPath expressions require variables or functions, your application may provide the class name of an XPathVariableResolver and XPathFunctionResolver. At type initialization time the engine instantiates the resolver instances and provides these to the XPathFactory.

This example shows the API to set this configuration.

ConfigurationEventTypeXMLDOM desc = new ConfigurationEventTypeXMLDOM();
desc.XPathFunctionResolver = typeof(MyXPathFunctionResolver).FullName;
desc.XPathVariableResolver = typeof(MyXPathVariableResolver).FullName;

11.4.3.5. Auto Fragment

This option is for use when a XSD schema is provided and determines whether the engine automatically creates an event type when a property expression transposes a property that is a complex type according to the schema.

An example:

ConfigurationEventTypeXMLDOM desc = new ConfigurationEventTypeXMLDOM();
desc.IsAutoFragment = false;

11.4.3.6. XPath Property Expression

By default Esper employs the built-in DOM walker implementation to evaluate XPath expressions, which is not namespace-aware.

This configuration setting, when set to true, instructs the engine to rewrite property expressions into XPath.

An example:

ConfigurationEventTypeXMLDOM desc = new ConfigurationEventTypeXMLDOM();
desc.IsXPathPropertyExpr = true;

11.4.3.7. Event Sender Setting

By default an EventSender for a given XML event type validates the root element name for which the type has been declared against the one provided by the System.Xml.XmlNode sent into the engine.

This configuration setting, when set to false, instructs an EventSender to not validate.

An example:

ConfigurationEventTypeXMLDOM desc = new ConfigurationEventTypeXMLDOM();
desc.IsEventSenderValidatesRoot = false;

11.4.4. Events represented by Plug-in Event Representations

As part of the extension API plug-in event representations allows an application to create new event types and event instances based on information available elsewhere. Please see Section 13.6, “Custom Event Representation” for details.

The configuration examples shown next use the configuration API to select settings. All options are also configurable via XML, please refer to the sample configuration XML in file esper.sample.cfg.xml.

11.4.4.1. Enabling an Custom Event Representation

Use the method AddPlugInEventRepresentation to enable a custom event representation, like this:

URI rootURI = new URI("type://mycompany/myproject/myname");
config.AddPlugInEventRepresentation(rootURI, 
    typeof(MyEventRepresentation).FullName, null);

The type:// part of the URI is an optional convention for the scheme part of an URI.

If your event representation takes initialization parameters, these are passed in as the last parameter. Initialization parameters can also be stored in the configuration XML, in which case they are passed as an XML string to the plug-in class.

11.4.4.2. Adding Plug-in Event Types

To register event types that your plug-in event representation creates in advance of creating an EPL statement (either at runtime or at configuration time), use the method AddPlugInEventType:

Uri childURI = new Uri("type://mycompany/myproject/myname/MyEvent");
configuration.AddPlugInEventType("MyEvent", new Uri[] {childURI}, null);

Your plug-in event type may take initialization parameters, these are passed in as the last parameter. Initialization parameters can also be stored in the configuration XML.

11.4.4.3. Setting Resolution URIs

The engine can invoke your plug-in event representation when an EPL statement is created with an event type name that the engine has not seen before. Plug-in event representations can resolve such names to an actual event type. In order to do this, you need to supply a list of resolution URIs. Use the PlugInEventTypeResolutionURIs property, at runtime or at configuration time:

Uri childURI = new Uri("type://mycompany/myproject/myname");
configuration.PlugInEventTypeNameResolutionURIs = new[] {childURI};

11.4.5. Class and package imports

Esper allows invocations of static library functions in expressions, as outlined in Section 8.1, “Single-row Function Reference”. This configuration item can be set to allow a partial rather than a fully qualified class name in such invocations. The imports work in the same way as in Java files, so both packages and classes can be imported.

select Math.Max(priceOne, PriceTwo)
// via configuration equivalent to
select System.Math.Max(priceOne, priceTwo)

Esper auto-imports the following library packages if no other configuration is supplied. This list is replaced with any configuration specified in a configuration file or through the API.

  • System

  • System.Collections

  • System.Text

In a XML configuration file the auto-import configuration may look as below:

<auto-import import-name="com.mycompany.mypackage"/>
<auto-import import-name="com.mycompany.myapp.MyUtilityClass"/>

Here is an example of providing imports via the API:

Configuration config = new Configuration();
config.AddImport("com.mycompany.mypackage");	// namespace import
config.AddImport("com.mycompany.mypackage.MyLib");   // type import

11.4.6. Cache Settings for From-Clause Method Invocations

Method invocations are allowed in the from clause in EPL, such that your application may join event streams to the data returned by a web service, or to data read from a distributed cache or object-oriented database, or obtain data by other means. A local cache may be placed in front of such method invocations through the configuration settings described herein.

The LRU cache is described in detail in Section 11.4.8.6.1, “LRU Cache”. The expiry-time cache documentation can be found in Section 11.4.8.6.2, “Expiry-time Cache”

The next XML snippet is a sample cache configuration that applies to methods provided by the classes 'MyFromClauseLookupLib' and 'MyFromClauseWebServiceLib'. The XML and API configuration understand both the fully-qualified CLR type name, as well as the simple class name:

<method-reference class-name="com.mycompany.MyFromClauseLookupLib">
  <expiry-time-cache max-age-seconds="10" purge-interval-seconds="10" ref-type="weak"/>
</method-reference> 	
<method-reference class-name="MyFromClauseWebServiceLib">
  <lru-cache size="1000"/>
</method-reference> 

11.4.7. Variables

Variables can be created dynamically in EPL via the create variable syntax but can also be configured at runtime and at configuration time.

A variable is declared by specifying a variable name, the variable type and an optional initialization value. The initialization value can be of the same or compatible type as the variable type, or can also be a String value that, when parsed, is compatible to the type declared for the variable.

In a XML configuration file the variable configuration may look as below. The Configuration API can also be used to configure variables.

<variable name="var_threshold" type="long" initialization-value="100"/>
<variable name="var_key" type="string"/>

Please find the list of valid values for the type attribute in Section 11.5, “Type Names”.

11.4.8. Relational Database Access

Esper has the capability to join event streams against historical data sources, such as a relational database. This section describes the configuration entries that the engine requires to access data stored in your database. Please see Section 4.15, “Accessing Relational Data via SQL” for information on the use of EPL queries that include historical data sources.

EPL queries that poll data from a relational database specify the name of the database as part of the EPL statement. The engine uses the configuration information described here to resolve the database name in the statement to database settings. The required and optional database settings are summarized below.

  • Database connections can be obtained via JDBC javax.xml.DataSource, via java.sql.DriverManager and via data source factory. Either one of these methods to obtain database connections is a required configuration.

  • Optionally, JDBC connection-level settings such as auto-commit, transaction isolation level, read-only and the catalog name can be defined.

  • Optionally, a connection lifecycle can be set to indicate to the engine whether the engine must retain connections or must obtain a new connection for each lookup and close the connection when the lookup is done (pooled).

  • Optionally, define a cache policy to allow the engine to retrieve data from a query cache, reducing the number of query executions.

Some of the settings can have important performance implications that need to be carefully considered in relationship to your database software, JDBC driver and runtime environment. This section attempts to outline such implications where appropriate.

The sample XML configuration file in the "etc" folder can be used as a template for configuring database settings. All settings are also available by means of the configuration API through the classes Configuration and ConfigurationDBRef.

11.4.8.1. Connections obtained via DataSource

This configuration causes Esper to obtain a database connection from a javax.sql.DataSource available from your JNDI provider.

The setting is most useful when running within an application server or when a JNDI directory is otherwise present in your Java VM. If your application environment does not provide an available DataSource, the next section outlines how to use Apache DBCP as a DataSource implementation with connection pooling options and outlines how to use a custom factory for DataSource implementations.

If your DataSource provides connections out of a connection pool, your configuration should set the collection lifecycle setting to pooled.

The snippet of XML below configures a database named mydb1 to obtain connections via a javax.sql.DataSource. The datasource-connection element instructs the engine to obtain new connections to the database mydb1 by performing a lookup via javax.naming.InitialContext for the given object lookup name. Optional environment properties for the InitialContext are also shown in the example.

<database-reference name="mydb1">
  <datasource-connection context-lookup-name="java:comp/env/jdbc/mydb">
    <env-property name="java.naming.factory.initial" value ="com.myclass.CtxFactory"/>
    <env-property name="java.naming.provider.url" value ="iiop://localhost:1050"/>
  </datasource-connection>
</database-reference>

To help you better understand how the engine uses this information to obtain connections, we have included the logic below.

if (envProperties.size() > 0) {
  initialContext = new InitialContext(envProperties);
}
else {
  initialContext = new InitialContext();
}
DataSource dataSource = (DataSource) initialContext.lookup(lookupName);
Connection connection = dataSource.getConnection();

In order to plug-in your own implementation of the DataSource interface, your application may use an existing JNDI provider as provided by an application server if running in a J2EE environment.

In case your application does not have an existing JNDI implementation to register a DataSource to provide connections, you may set the java.naming.factory.initial property in the configuration to point to your application's own implementation of the javax.naming.spi.InitialContextFactory interface that can return your application DataSource though the javax.naming.Context provided by the factory implementation. Please see Java Naming and Directory Interface (JNDI) API documentation for further information.

11.4.8.2. Connections obtained via DataSource Factory

This section describes how to use Apache Commons Database Connection Pooling (Apache DBCP) with Esper. We also explain how to provide a custom application-specific DataSource factory if not using Apache DBCP.

If your DataSource provides connections out of a connection pool, your configuration should set the collection lifecycle setting to pooled.

Apache DBCP provides comprehensive means to test for dead connections or grow and shrik a connection pool. Configuration properties for Apache DBCP can be found at Apache DBCP configuration. The listed properties are passed to Apache DBCP via the properties list provided as part of the Esper configuration.

The snippet of XML below is an example that configures a database named mydb3 to obtain connections via the pooling DataSource provided by Apache DBCP BasicDataSourceFactory.

The listed properties are passed to DBCP to instruct DBCP how to manage the connection pool. The settings below initialize the connection pool to 2 connections and provide the validation query select 1 from dual for DBCP to validate a connection before providing a connection from the pool to Esper:

<database-reference name="mydb3">
  <!-- For a complete list of properties see Apache DBCP. -->
  <datasourcefactory-connection class-name="org.apache.commons.dbcp.BasicDataSourceFactory">	
    <env-property name="username" value ="myusername"/>
    <env-property name="password" value ="mypassword"/>
    <env-property name="driverClassName" value ="com.mysql.jdbc.Driver"/>
    <env-property name="url" value ="jdbc:mysql://localhost/test"/>
    <env-property name="initialSize" value ="2"/>
    <env-property name="validationQuery" value ="select 1 from dual"/>
  </datasourcefactory-connection>
  <connection-lifecycle value="pooled"/>
</database-reference>

The same configuration options provided through the API:

Properties props = new Properties();
props.put("username", "myusername");
props.put("password", "mypassword");
props.put("driverClassName", "com.mysql.jdbc.Driver");
props.put("url", "jdbc:mysql://localhost/test");
props.put("initialSize", 2);
props.put("validationQuery", "select 1 from dual");

ConfigurationDBRef configDB = new ConfigurationDBRef();
// BasicDataSourceFactory is an Apache DBCP import
configDB.setDataSourceFactory(props, BasicDataSourceFactory.class.getName());
configDB.setConnectionLifecycleEnum(ConfigurationDBRef.ConnectionLifecycleEnum.POOLED);

Configuration configuration = new Configuration();;
configuration.addDatabaseReference("mydb3", configDB);

Apache Commons DBCP is a separate download and not provided as part of the Esper distribution. The Apache Commons DBCP jar file requires the Apache Commons Pool jar file.

Your application can provide its own factory implementation for DataSource instances: Set the class name to the name of the application class that provides a public static method named createDataSource which takes a single Properties object as parameter and returns a DataSource implementation. For example:

configDB.setDataSourceFactory(props, MyOwnDataSourceFactory.class.getName());
...
class MyOwnDataSourceFactory {
  public static DataSource CreateDataSource(Properties properties) {
    return new MyDataSourceImpl(properties);
  }
}

11.4.8.3. Connections obtained via DriverManager

The next snippet of XML configures a database named mydb2 to obtain connections via java.sql.DriverManager. The drivermanager-connection element instructs the engine to obtain new connections to the database mydb2 by means of Class.forName and DriverManager.getConnection using the class name, URL and optional username, password and connection arguments.

<database-reference name="mydb2">
  <drivermanager-connection class-name="my.sql.Driver" 
        url="jdbc:mysql://localhost/test?user=root&amp;password=mypassword" 
        user="myuser" password="mypassword">
    <connection-arg name="user" value ="myuser"/>
    <connection-arg name="password" value ="mypassword"/>
    <connection-arg name="somearg" value ="someargvalue"/>
  </drivermanager-connection>
</database-reference>

The username and password are shown in multiple places in the XML only as an example. Please check with your database software on the required information in URL and connection arguments.

11.4.8.4. Connections-level settings

Additional connection-level settings can optionally be provided to the engine which the engine will apply to new connections. When the engine obtains a new connection, it applies only those settings to the connection that are explicitly configured. The engine leaves all other connection settings at default values.

The below XML is a sample of all available configuration settings. Please refer to the Java API JavaDocs for System.Data.DbConnection for more information to each option or check the documentation of your driver and database software.

<database-reference name="mydb2">
... configure data source or driver manager settings...
  <connection-settings auto-commit="true" catalog="mycatalog" 
      read-only="true" transaction-isolation="1" />
</database-reference>

The read-only setting can be used to indicate to your database engine that SQL statements are read-only. The transaction-isolation and auto-commit help you database software perform the right level of locking and lock release. Consider setting these values to reduce transactional overhead in your database queries.

11.4.8.5. Connections lifecycle settings

By default the engine retains a separate database connection for each started EPL statement. However, it is possible to override this behavior and require the engine to obtain a new database connection for each lookup, and to close that database connection after the lookup is completed. This often makes sense when you have a large number of EPL statements and require pooling of connections via a connection pool.

In the pooled setting, the engine obtains a database connection from the data source or driver manager for every query, and closes the connection when done, returning the database connection to the pool if using a pooling data source.

In the retain setting, the engine retains a separate dedicated database connection for each statement and does not close the connection between uses.

The XML for this option is below. The connection lifecycle allows the following values: pooled and retain.

<database-reference name="mydb2">
... configure data source or driver manager settings...
    <connection-lifecycle value="pooled"/>
</database-reference>

11.4.8.6. Cache settings

Cache settings can dramatically reduce the number of database queries that the engine executes for EPL statements. If no cache setting is specified, the engine does not cache query results and executes a separate database query for every event.

Caches store the results of database queries and make these results available to subsequent queries using the exact same query parameters as the query for which the result was stored. If your query returns one or more rows, the cache keep the result rows of the query keyed to the parameters of the query. If your query returns no rows, the cache also keeps the empty result. Query results are held by a cache until the cache entry is evicted. The strategies available for evicting cached query results are listed next.

11.4.8.6.1. LRU Cache

The least-recently-used (LRU) cache is configured by a maximum size. The cache discards the least recently used query results first once the cache reaches the maximum size.

The XML configuration entry for a LRU cache is as below. This entry configures an LRU cache holding up to 1000 query results.

<database-reference name="mydb">
... configure data source or driver manager settings...
    <lru-cache size="1000"/>
</database-reference>
11.4.8.6.2. Expiry-time Cache

The expiry time cache is configured by a maximum age in seconds, a purge interval and an optional reference type. The cache discards (on the get operation) any query results that are older then the maximum age so that stale data is not used. If the cache is not empty, then every purge interval number of seconds the engine purges any expired entries from the cache.

The XML configuration entry for an expiry-time cache is as follows. The example configures an expiry time cache in which prior query results are valid for 60 seconds and which the engine inspects every 2 minutes to remove query results older then 60 seconds.

<database-reference name="mydb">
... configure data source or driver manager settings...
    <expiry-time-cache max-age-seconds="60" purge-interval-seconds="120" />
</database-reference>

By default, the expiry-time cache is backed by a java.util.WeakHashMap and thus relies on weak references. That means that cached SQL results can be freed during garbage collection.

Via XML or using the configuration API the type of reference can be configured to not allow entries to be garbage collected, by setting the ref-type property to hard:

<database-reference name="mydb">
... configure data source or driver manager settings...
    <expiry-time-cache max-age-seconds="60" purge-interval-seconds="120" ref-type="hard"/>
</database-reference>

The last setting for the cache reference type is soft: This strategy allows the garbage collection of cache entries only when all other weak references have been collected.

11.4.8.7. Column Change Case

This setting instructs the engine to convert to lower- or uppercase any output column names returned by your database system. When using Oracle relational database software, for example, column names can be changed to lowercase via this setting.

A sample XML configuration entry for this setting is:

<column-change-case value="lowercase"/>

11.4.8.8. SQL Types Mapping

By providing a mapping of SQL types (java.sql.Types) to Java built-in types your code can avoid using sometimes awkward default database types and can easily change the way Esper returns Java types for columns returned by a SQL query.

The mapping maps a constant as defined by java.sql.Types to a Java built-in type of any of the following Java type names: String, BigDecimal, Boolean, Byte, Short, Int, Long, Float, Double, ByteArray, SqlDate, SqlTime, SqlTimestamp. The Java type names are not case-sensitive.

A sample XML configuration entry for this setting is shown next. The sample maps Types.NUMERIC which is a constant value of 2 per JDBC API to the Java int type.

<sql-types-mapping sql-type="2" java-type="int" />

11.4.8.9. Metadata Origin

This setting controls how the engine retrieves SQL statement metadata from JDBC prepared statements.

Table 11.3. Syntax and results of aggregate functions

OptionDescription
default

By default, the engine detects the driver name and queries prepared statement metadata if the driver is not an Oracle database driver. For Oracle drivers, the engine uses lexical analysis of the SQL statement to construct a sample SQL statement and then fires that statement to retrieve statement metadata.

metadata

The engine always queries prepared statement metadata regardless of the database driver used.

sample

The engine always uses lexical analysis of the SQL statement to construct a sample SQL statement, and then fires that statement to retrieve statement metadata.

11.4.9. Engine Settings related to Concurrency and Threading

11.4.9.1. Preserving the order of events delivered to listeners

In multithreaded environments, this setting controls whether dispatches of statement result events to listeners preserve the ordering in which a statement processes events. By default the engine guarantees that it delivers a statement's result events to statement listeners in the order in which the result is generated. This behavior can be turned off via configuration as below.

The next code snippet shows how to control this feature:

Configuration config = new Configuration();
config.EngineDefaults.Threading.IsListenerDispatchPreserveOrder = false;
engine = EPServiceProviderManager.GetDefaultProvider(config);

And the XML configuration file can also control this feature by adding the following elements:

<engine-settings>
  <defaults>
    <threading>
      <listener-dispatch preserve-order="true" timeout-msec="1000" locking="spin"/>
    </threading>
  </defaults>
</engine-settings>

As discussed, by default the engine can temporarily block another processing thread when delivering result events to listeners in order to preserve the order in which results are delivered to a given statement. The maximum time the engine blocks a thread can also be configured, and by default is set to 1 second.

As such delivery locks are typically held for a very short amount of time, the default blocking technique employs a spin lock (There are two techniques for implementing blocking; having the operating system suspend the thread until it is awakened later or using spin locks). While spin locks are CPU-intensive and appear inefficient, a spin lock can be more efficient than suspending the thread and subsequently waking it up, especially if the lock in question is held for a very short time. That is because there is significant overhead to suspending and rescheduling a thread.

The locking technique can be changed to use a blocking strategy that suspends the thread, by means of setting the locking property to 'suspend'.

11.4.9.2. Preserving the order of events for insert-into streams

In multithreaded environments, this setting controls whether statements producing events for other statements via insert-into preserve the order of delivery within the producing and consuming statements, allowing statements that consume other statement's events to behave deterministic in multithreaded applications, if the consuming statement requires such determinism. By default, the engine makes this guarantee (the setting is on).

Take, for example, an application where a single statement (S1) inserts events into a stream that another statement (S2) further evaluates. A multithreaded application may have multiple threads processing events into statement S1. As statement S1 produces events for consumption by statement S2, such results may need to be delivered in the exact order produced as the consuming statement may rely on the order received. For example, if the first statement counts the number of events, the second statement may employ a pattern that inspects counts and thus expect the counts posted by statement S1 to continuously increase by 1 even though multiple threads process events.

The engine may need to block a thread such that order of delivery is maintained, and statements that require order (such as pattern detection, previous and prior functions) receive a deterministic order of events. The settings available control the blocking technique and parameters. As described in the section immediately prior, the default blocking technique employs spin locks per statement inserting events for consumption, as the locks in questions are typically held a very short time. The 'suspend' blocking technique can be configured and a timeout value can also defined.

The XML configuration file may change settings via the following elements:

<engine-settings>
  <defaults>
    <threading>
      <insert-into-dispatch preserve-order="true" timeout-msec="100" locking="spin"/>
    </threading>
  </defaults>
</engine-settings>

11.4.9.3. Internal Timer Settings

This option can be used to disable the internal timer thread and such have the application supply external time events, as well as to set a timer resolution.

The next code snippet shows how to disable the internal timer thread via the configuration API:

Configuration config = new Configuration();
  config.EngineDefaults.Threading.UsInternalTimerEnabled = false;

This snippet of XML configuration leaves the internal timer enabled (the default) and sets a resolution of 200 milliseconds (the default is 100 milliseconds):

<engine-settings>
  <defaults>
    <threading>
      <internal-timer enabled="true" msec-resolution="200"/>
    </threading>
  </defaults>
</engine-settings>

We recommend that when disabling the internal timer, applications send an external timer event setting the start time before creating statements, such that statement start time is well-defined.

11.4.9.4. Advanced Threading Options

The settings described herein are for enabling advanced threading options for inbound, outbound, timer and route executions.

Take the next snippet of XML configuration as an example. It configures all threading options to 2 threads, which may not be suitable to your application, however demonstrates the configuration:

<engine-settings>
  <defaults>
    <threading>
      <threadpool-inbound enabled="true" num-threads="2"/>
      <threadpool-outbound enabled="true" num-threads="2" capacity="1000"/>
      <threadpool-timerexec enabled="true" num-threads="2"/>
      <threadpool-routeexec enabled="true" num-threads="2"/>
    </threading>
  </defaults>
</engine-settings>

By default, queues are unbound and backed by java.util.concurrent.LinkedBlockingQueue. The optional capacity attribute can be set to instruct the threading option to configure a capacity-bound queue with a sender-wait (blocking put) policy, backed ArrayBlockingQueue.

This example uses the API for configuring inbound threading :

Configuration config = new Configuration();
config.EngineDefaults.Threading.IsThreadPoolInbound = true;
config.EngineDefaults.Threading.ThreadPoolInboundNumThreads = 2;

With a bounded work queue, the queue size and pool size should be tuned together. A large queue coupled with a small pool can help reduce memory usage, CPU usage, and context switching, at the cost of potentially constraining throughput.

11.4.10. Engine Settings related to Event Metadata

11.4.10.1. Type Property Names, Case Sensitivity and Accessor Style

The engine-wide settings discussed here are used when you want to control case sensitivity or accessor style for all event classes as a default. The two settings are found under class-property-resolution under event-meta in the XML configuration.

To control the case sensitivity as discussed in Section 11.4.1.6, “Case Sensitivity and Property Names”, add the style attribute in the XML configuration to set a default case sensitivity applicable to all event classes unless specifically overridden by class-specific configuration. The default case sensitivity is case_sensitive (case sensitivity turned on).

To control the accessor style as discussed in Section 11.4.1.3, “Legacy Event Classes”, add the accessor-style attribute in the XML configuration to set a default accessor style applicable to all event classes unless specifically overridden by class-specific configuration. The default accessor style is native accessor style.

The next code snippet shows how to control this feature via the API:

Configuration config = new Configuration();
config.EngineDefaults.EventMeta.ClassPropertyResolutionStyle =  PropertyResolutionStyle.CASE_INSENSITIVE;
config.EngineDefaults.EventMeta.DefaultAccessorStyle = AccessorStyleEnum.PUBLIC;

11.4.11. Engine Settings related to View Resources

11.4.11.1. Sharing View Resources between Statements

The engine by default attempts to optimize resource usage and thus re-uses or shares views between statements that declare same views. However, in multi-threaded environments, this can lead to reduced concurrency as locking for shared view resources must take place. Via this setting this behavior can be turned off for higher concurrency in multi-threaded processing.

The next code snippet outlines the API to turn off view resource sharing between statements:

Configuration config = new Configuration();
config.EngineDefaults.ViewResources.ShareViews = false;

11.4.11.2. Configuring Multi-Expiry Policy Defaults

By default