org.apache.beam.sdk.transforms

Class DoFn.Context

java.lang.Object

org.apache.beam.sdk.transforms.DoFn.Context

Direct Known Subclasses:

DoFn.ProcessContext

Enclosing class:

DoFn<InputT,OutputT>

public abstract class DoFn.Context
extends Object

Information accessible to all methods in this DoFn. Used primarily to output elements.

Constructor Summary

Constructors

Constructor and Description
Context()

Method Summary

Modifier and Type	Method and Description
protected abstract <AggInputT,AggOutputT> Aggregator<AggInputT,AggOutputT>	createAggregatorInternal(String name, Combine.CombineFn<AggInputT,?,AggOutputT> combiner) Creates an Aggregator in the DoFn context with the specified name and aggregation logic specified by Combine.CombineFn.
abstract PipelineOptions	getPipelineOptions() Returns the PipelineOptions specified with the PipelineRunner invoking this DoFn.
abstract void	output(OutputT output) Adds the given element to the main output PCollection.
abstract void	outputWithTimestamp(OutputT output, Instant timestamp) Adds the given element to the main output PCollection, with the given timestamp.
protected void	setupDelegateAggregators() Sets up Aggregators created by the DoFn so they are usable within this context.
abstract <T> void	sideOutput(TupleTag<T> tag, T output) Adds the given element to the side output PCollection with the given tag.
abstract <T> void	sideOutputWithTimestamp(TupleTag<T> tag, T output, Instant timestamp) Adds the given element to the specified side output PCollection, with the given timestamp.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Context

public Context()

Method Detail

getPipelineOptions

public abstract PipelineOptions getPipelineOptions()

Returns the PipelineOptions specified with the PipelineRunner invoking this DoFn. The PipelineOptions will be the default running via DoFnTester.

output

public abstract void output(OutputT output)

Adds the given element to the main output PCollection.

Once passed to output the element should be considered immutable and not be modified in any way. It may be cached or retained by the Dataflow runtime or later steps in the pipeline, or used in other unspecified ways.

If invoked from processElement, the output element will have the same timestamp and be in the same windows as the input element passed to processElement.

If invoked from startBundle or finishBundle, this will attempt to use the WindowFn of the input PCollection to determine what windows the element should be in, throwing an exception if the WindowFn attempts to access any information about the input element. The output element will have a timestamp of negative infinity.

outputWithTimestamp

public abstract void outputWithTimestamp(OutputT output,
                                         Instant timestamp)

Adds the given element to the main output PCollection, with the given timestamp.

Once passed to outputWithTimestamp the element should not be modified in any way.

If invoked from processElement, the timestamp must not be older than the input element's timestamp minus getAllowedTimestampSkew. The output element will be in the same windows as the input element.

If invoked from startBundle or finishBundle, this will attempt to use the WindowFn of the input PCollection to determine what windows the element should be in, throwing an exception if the WindowFn attempts to access any information about the input element except for the timestamp.

sideOutput

public abstract <T> void sideOutput(TupleTag<T> tag,
                                    T output)

Adds the given element to the side output PCollection with the given tag.

Once passed to sideOutput the element should not be modified in any way.

The caller of ParDo uses withOutputTags to specify the tags of side outputs that it consumes. Non-consumed side outputs, e.g., outputs for monitoring purposes only, don't necessarily need to be specified.

The output element will have the same timestamp and be in the same windows as the input element passed to processElement.

If invoked from startBundle or finishBundle, this will attempt to use the WindowFn of the input PCollection to determine what windows the element should be in, throwing an exception if the WindowFn attempts to access any information about the input element. The output element will have a timestamp of negative infinity.

See Also:

ParDo.withOutputTags(org.apache.beam.sdk.values.TupleTag<OutputT>, org.apache.beam.sdk.values.TupleTagList)

sideOutputWithTimestamp

public abstract <T> void sideOutputWithTimestamp(TupleTag<T> tag,
                                                 T output,
                                                 Instant timestamp)

Adds the given element to the specified side output PCollection, with the given timestamp.

Once passed to sideOutputWithTimestamp the element should not be modified in any way.

If invoked from processElement, the timestamp must not be older than the input element's timestamp minus getAllowedTimestampSkew. The output element will be in the same windows as the input element.

If invoked from startBundle or finishBundle, this will attempt to use the WindowFn of the input PCollection to determine what windows the element should be in, throwing an exception if the WindowFn attempts to access any information about the input element except for the timestamp.

See Also:

ParDo.withOutputTags(org.apache.beam.sdk.values.TupleTag<OutputT>, org.apache.beam.sdk.values.TupleTagList)

createAggregatorInternal

@Experimental(value=AGGREGATOR)
protected abstract <AggInputT,AggOutputT> Aggregator<AggInputT,AggOutputT> createAggregatorInternal(String name,
                                                                                                                                    Combine.CombineFn<AggInputT,?,AggOutputT> combiner)

Creates an Aggregator in the DoFn context with the specified name and aggregation logic specified by Combine.CombineFn.

For internal use only.

Parameters:

name - the name of the aggregator

combiner - the Combine.CombineFn to use in the aggregator

Returns:

an aggregator for the provided name and Combine.CombineFn in this context

setupDelegateAggregators

@Experimental(value=AGGREGATOR)
protected final void setupDelegateAggregators()

Sets up Aggregators created by the DoFn so they are usable within this context.

This method should be called by runners before DoFn.startBundle(org.apache.beam.sdk.transforms.DoFn<InputT, OutputT>.Context) is executed.