com.google.cloud.dataflow.sdk.io

Class Sink.WriteOperation<T,WR>

java.lang.Object

com.google.cloud.dataflow.sdk.io.Sink.WriteOperation<T,WR>

Type Parameters:

T - The type of objects to write

WR - The result of a per-bundle write

All Implemented Interfaces:

java.io.Serializable

Direct Known Subclasses:

FileBasedSink.FileBasedWriteOperation

Enclosing class:

Sink<T>

public abstract static class Sink.WriteOperation<T,WR>
extends java.lang.Object
implements java.io.Serializable

A WriteOperation defines the process of a parallel write of objects to a Sink.

The WriteOperation defines how to perform initialization and finalization of a parallel write to a sink as well as how to create a Sink.Writer object that can write a bundle to the sink.

Since operations in Dataflow may be run multiple times for redundancy or fault-tolerance, the initialization and finalization defined by a WriteOperation must be idempotent.

WriteOperations may be mutable; a WriteOperation is serialized after the call to initialize method and deserialized before calls to createWriter and finalized. However, it is not reserialized after createWriter, so createWriter should not mutate the state of the WriteOperation.

See Sink for more detailed documentation about the process of writing to a Sink.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
Sink.WriteOperation()

Method Summary

Methods

Modifier and Type	Method and Description
abstract Sink.Writer<T,WR>	createWriter(PipelineOptions options) Creates a new Sink.Writer to write a bundle of the input to the sink.
abstract void	finalize(java.lang.Iterable<WR> writerResults, PipelineOptions options) Given an Iterable of results from bundle writes, performs finalization after writing and closes the sink.
abstract Sink<T>	getSink() Returns the Sink that this write operation writes to.
Coder<WR>	getWriterResultCoder() Returns a coder for the writer result type.
abstract void	initialize(PipelineOptions options) Performs initialization before writing to the sink.

Methods inherited from class java.lang.Object

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

Constructor Detail

Sink.WriteOperation

public Sink.WriteOperation()

Method Detail

initialize

public abstract void initialize(PipelineOptions options)
                         throws java.lang.Exception

Performs initialization before writing to the sink. Called before writing begins.

Throws:

java.lang.Exception

finalize

public abstract void finalize(java.lang.Iterable<WR> writerResults,
            PipelineOptions options)
                       throws java.lang.Exception

Given an Iterable of results from bundle writes, performs finalization after writing and closes the sink. Called after all bundle writes are complete.

The results that are passed to finalize are those returned by bundles that completed successfully. Although bundles may have been run multiple times (for fault-tolerance), only one writer result will be passed to finalize for each bundle. An implementation of finalize should perform clean up of any failed and successfully retried bundles. Note that these failed bundles will not have their writer result passed to finalize, so finalize should be capable of locating any temporary/partial output written by failed bundles.

A best practice is to make finalize atomic. If this is impossible given the semantics of the sink, finalize should be idempotent, as it may be called multiple times in the case of failure/retry or for redundancy.

Note that the iteration order of the writer results is not guaranteed to be consistent if finalize is called multiple times.

Parameters:

writerResults - an Iterable of results from successful bundle writes.

Throws:

java.lang.Exception

createWriter

public abstract Sink.Writer<T,WR> createWriter(PipelineOptions options)
                                        throws java.lang.Exception

Creates a new Sink.Writer to write a bundle of the input to the sink.

The bundle id that the writer will use to uniquely identify its output will be passed to Sink.Writer.open(java.lang.String).

Must not mutate the state of the WriteOperation.

Throws:

java.lang.Exception

getSink

public abstract Sink<T> getSink()

Returns the Sink that this write operation writes to.

getWriterResultCoder

public Coder<WR> getWriterResultCoder()

Returns a coder for the writer result type.