Event properties to be populated can simply be added to the parameters.

If your declaration provides an event type for the output stream then BeaconSource will populate event properties of the underlying events. If no event type is specified, BeaconSource creates an anonymous object-array event type to carry the event properties that are generated and associates this type with its output stream.

Examples are:

create dataflow MyDataFlow
  create schema SampleSchema(tagId string, locX double),	// sample type			
			
  // BeaconSource that produces empty object-array events without delay 
  // or interval until cancelled.
  BeaconSource -> stream.one {}
  
  // BeaconSource that produces one RFIDSchema event populating event properties
  // from a user-defined function "generateTagId" and the provided values.
  BeaconSource -> stream.two<SampleSchema> {
    iterations : 1,
    tagId : generateTagId(),
    locX : 10
  }
  
  // BeaconSource that produces 10 object-array events populating
  // the price property with a random value.
  BeaconSource -> stream.three {
    iterations : 10,
    interval : 10, // every 10 seconds
    initialDelay : 5, // start after 5 seconds
    price : Math.random() * 100
  }

If a statement name is provided, the operator subscribes to output events of the statement if the statement exists or when it gets created at a later point in time.

If a statement filter is provided instead, the operator subscribes to output events of all statements that currently exist and pass the filter pass method or that get created at a later point in time and pass the filter pass method.

The collector can be specified to transform output events. If no collector is specified the operator submits the underlying events of the insert stream received from the statement. The collector object must implement the interface EPDataFlowIRStreamCollector.

Examples are:

create dataflow MyDataFlow
  create schema SampleSchema(tagId string, locX double),	// sample type			
			
  // Consider only the statement named MySelectStatement when it exists.
  // No transformation.
  EPStatementSource -> stream.one<eventbean<?>> {
    statementName : 'MySelectStatement'
  }
  
  // Consider all statements that match the filter object provided.
  // No transformation.
  EPStatementSource -> stream.two<eventbean<?>> {
    statementFilter : {
      class : 'com.mycompany.filters.MyStatementFilter'
    }
  }
  
  // Consider all statements that match the filter object provided.
  // With collector that performs transformation.
  EPStatementSource -> stream.two<SampleSchema> {
    collector : {
      class : 'com.mycompany.filters.MyCollector'
    },
    statementFilter : {
      class : 'com.mycompany.filters.MyStatementFilter'
    }
  }

The collector can be specified to transform data flow events to event bus events. If no collector is specified the operator submits the events directly to the event bus. The collector object must implement the interface EPDataFlowEventCollector.

Examples are:

create dataflow MyDataFlow
  BeaconSource -> instream<SampleSchema> {}  // produces a sample stream
  
  // Send SampleSchema events produced by beacon to the event bus.
  EventBusSink(instream) {}
  
  // Send SampleSchema events produced by beacon to the event bus.
  // With collector that performs transformation.
  EventBusSink(instream) {
    collector : {
      class : 'com.mycompany.filters.MyCollector'
    }
  }

The collector can be specified to transform output events. If no collector is specified the operator submits the underlying events of the stream received from the event bus. The collector object must implement the interface EPDataFlowEventBeanCollector.

The filter is an expression that the event bus compiles and efficiently matches even in the presence of a large number of event bus sources. The filter expression must return a boolean-typed value, returning true for those events that the event bus passes to the operator.

Examples are:

create dataflow MyDataFlow

  // Receive all SampleSchema events from the event bus.
  // No transformation.
  EventBusSource -> stream.one<SampleSchema> {}
  
  // Receive all SampleSchema events with tag id '001' from the event bus.
  // No transformation.
  EventBusSource -> stream.one<SampleSchema> {
    filter : tagId = '001'
  }

  // Receive all SampleSchema events from the event bus.
  // With collector that performs transformation.
  EventBusSource -> stream.two<SampleSchema> {
    collector : {
      class : 'com.mycompany.filters.MyCollector'
    },
  }

Examples are:

create dataflow MyDataFlow
  create schema SampleSchema(tagId string, locX double),	// sample type
  BeaconSource -> samplestream<SampleSchema> {}  // sample source
  
  // Filter all events that have a tag id of '001'
  Filter(samplestream) -> tags_001 {
    filter : tagId = '001' 
  }
  
  // Filter all events that have a tag id of '001', 
  // putting all other events into the second stream
  Filter(samplestream) -> tags_001, tags_other {
    filter : tagId = '001' 
  }

Examples are:

create dataflow MyDataFlow
  BeaconSource -> instream {}  // produces sample stream to use below
  
  // Output textual event to log using defaults.
  LogSink(instream) {}
  
  // Output JSON-formatted to console.
  LogSink(instream) {
    format : 'json',
    layout : '%t [%e]',
    log : false,
    linefeed : true,
    title : 'My Custom Title:'
  }

Set the optional iterate flag to false (the default) to have the operator output results continuously. Set the iterate flag to true to indicate that the operator outputs results only when the final marker arrives. If iterate is true then output rate limiting clauses are not supported.

The select parameter is required and provides an EPL select statement within parenthesis. For each input port the statement should list the input stream name or the alias name in the from clause. Only filter-based streams are allowed in the from clause and patterns or named windows are not supported. Also not allowed are the insert into clause, the irstream keyword and subselects.

The Select operator determines the event type of output events based on the select clause. It is not necessary to declare an event type for the output stream.

Examples are:

create dataflow MyDataFlow
  create schema SampleSchema(tagId string, locX double),	// sample type			
  BeaconSource -> instream<SampleSchema> {}  // sample stream
  BeaconSource -> secondstream<SampleSchema> {}  // sample stream
  
  // Simple continuous count of events
  Select(instream) -> outstream {
    select: (select count(*) from instream)
  }
  
  // Demonstrate use of alias
  Select(instream as myalias) -> outstream {
    select: (select count(*) from myalias)
  }
  
  // Output only when the final marker arrives
  Select(instream as myalias) -> outstream {
    select: (select count(*) from myalias),
    iterate: true
  }

  // Same input port for the two sample streams
  Select( (instream, secondstream) as myalias) -> outstream {
    select: (select count(*) from myalias)
  }

  // A join with multiple input streams,
  // joining the last event per stream forming pairs
  Select(instream, secondstream) -> outstream {
    select: (select a.tagId, b.tagId 
        from instream#lastevent as a, secondstream#lastevent as b)
  }
  
  // A join with multiple input streams and using aliases.
  Select(instream as S1, secondstream as S2) -> outstream {
    select: (select a.tagId, b.tagId 
        from S1#lastevent as a, S2#lastevent as b)
  }