Deprecated API

Contents

• Interfaces
• Classes
• Enums
• Fields
• Methods

Interfaces

Interface	Description
org.apache.beam.sdk.testing.StreamingIT	tests which use unbounded PCollections should be in the category UsesUnboundedPCollections. Beyond that, it is up to the runner and test configuration to decide whether to run in streaming mode.

Classes

Class	Description
org.apache.beam.sdk.coders.Coder.Context	To implement a coder, do not use any Coder.Context. Just implement only those abstract methods which do not accept a Coder.Context and leave the default implementations for methods accepting a Coder.Context.
org.apache.beam.sdk.io.TextIO.ReadAll	See TextIO.readAll() for details.
org.apache.beam.sdk.schemas.GetterBasedSchemaProvider	new implementations should extend the GetterBasedSchemaProviderV2 class' methods which receive TypeDescriptors instead of ordinary Classes as arguments, which permits to support generic type signatures during schema inference
org.apache.beam.sdk.transforms.ApproximateUnique	Consider using ApproximateCountDistinct in the zetasketch extension module, which makes use of the HllCount implementation. If ApproximateCountDistinct does not meet your needs then you can directly use HllCount. Direct usage will also give you access to save intermediate aggregation result into a sketch for later processing. For example, to estimate the number of distinct elements in a PCollection<String>: PCollection<String> input = ...; PCollection<Long> countDistinct = input.apply(HllCount.Init.forStrings().globally()).apply(HllCount.Extract.globally()); For more details about using HllCount and the zetasketch extension module, see https://s.apache.org/hll-in-beam#bookmark=id.v6chsij1ixo7.
org.apache.beam.sdk.transforms.Combine.SimpleCombineFn
org.apache.beam.sdk.transforms.DoFnTester	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.Top.Largest	use Top.Natural instead
org.apache.beam.sdk.transforms.Top.Smallest	use Top.Reversed instead
org.apache.beam.sdk.util.BitSetCoder	use BitSetCoder instead
org.apache.beam.sdk.util.construction.CreatePCollectionViewTranslation	this should generally be done as part of ParDo translation, or moved into a dedicated runners-core-construction auxiliary class
org.apache.beam.sdk.util.construction.CreatePCollectionViewTranslation.Registrar	runners should move away from translating `CreatePCollectionView` and treat this as part of the translation for a `ParDo` side input.
org.apache.beam.sdk.values.PCollectionViews.IterableViewFn	See PCollectionViews.IterableViewFn2.
org.apache.beam.sdk.values.PCollectionViews.ListViewFn	See PCollectionViews.ListViewFn2.
org.apache.beam.sdk.values.PCollectionViews.MapViewFn	See PCollectionViews.MapViewFn2.
org.apache.beam.sdk.values.PCollectionViews.MultimapViewFn	See PCollectionViews.MultimapViewFn2.
org.apache.beam.sdk.values.PCollectionViews.SingletonViewFn	See PCollectionViews.SingletonViewFn2.
org.apache.beam.sdk.values.ShardedKey	Use ShardedKey instead.
org.apache.beam.sdk.values.WindowedValues.ValueOnlyWindowedValueCoder	Use ParamWindowedValueCoder instead, it is a general purpose implementation of the same concept but makes timestamp, windows and pane info configurable.

Enums

Enum	Description
org.apache.beam.sdk.io.CompressedSource.CompressionMode	Use Compression instead
org.apache.beam.sdk.io.FileBasedSink.CompressionType	use Compression.
org.apache.beam.sdk.io.TextIO.CompressionType	Use Compression.
org.apache.beam.sdk.io.TFRecordIO.CompressionType	Use Compression.
org.apache.beam.sdk.transforms.DoFnTester.CloningBehavior	Use TestPipeline with the DirectRunner.

Fields

Field	Description
org.apache.beam.sdk.schemas.parser.generated.FieldSpecifierNotationLexer.tokenNames	Use FieldSpecifierNotationLexer.VOCABULARY instead.
org.apache.beam.sdk.schemas.parser.generated.FieldSpecifierNotationParser.tokenNames	Use FieldSpecifierNotationParser.VOCABULARY instead.
org.apache.beam.sdk.util.construction.PTransformTranslation.CREATE_VIEW_TRANSFORM_URN	runners should move away from translating `CreatePCollectionView` and treat this as part of the translation for a `ParDo` side input.
org.apache.beam.sdk.util.construction.PTransformTranslation.READ_TRANSFORM_URN	SDKs should move away from creating `Read` transforms and migrate to using Impulse + SplittableDoFns.
org.apache.beam.sdk.util.construction.PTransformTranslation.SPLITTABLE_PROCESS_ELEMENTS_URN	runners should move away from using `SplittableProcessElements` and prefer to internalize any necessary SplittableDoFn expansion.
org.apache.beam.sdk.util.construction.PTransformTranslation.SPLITTABLE_PROCESS_KEYED_URN	runners should move away from using `SplittableProcessKeyedElements` and prefer to internalize any necessary SplittableDoFn expansion.

Methods

Method	Description
org.apache.beam.sdk.coders.Coder.decode(InputStream, Coder.Context)	only implement and call Coder.decode(InputStream)
org.apache.beam.sdk.coders.Coder.encode(T, OutputStream, Coder.Context)	only implement and call Coder.encode(Object value, OutputStream)
org.apache.beam.sdk.coders.CoderRegistry.getCoder(TypeDescriptor<OutputT>, TypeDescriptor<InputT>, Coder<InputT>)	This method is to change in an unknown backwards incompatible way once support for this functionality is refined.
org.apache.beam.sdk.coders.CoderRegistry.getOutputCoder(SerializableFunction<InputT, OutputT>, Coder<InputT>)	This method is to change in an unknown backwards incompatible way once support for this functionality is refined.
org.apache.beam.sdk.expansion.ExternalTransformRegistrar.knownBuilders()	Prefer implementing 'knownBuilderInstances'. This method will be removed in a future version of Beam.
org.apache.beam.sdk.fn.data.CloseableFnDataReceiver.flush()	to be removed once splitting/checkpointing are available in SDKs and rewinding in readers.
org.apache.beam.sdk.fn.server.GrpcFnServer.create(ServiceT, Endpoints.ApiServiceDescriptor)	This create function is used for Dataflow migration purpose only.
org.apache.beam.sdk.io.CountingSource.unbounded()	use GenerateSequence instead
org.apache.beam.sdk.io.CountingSource.unboundedWithTimestampFn(SerializableFunction<Long, Instant>)	use GenerateSequence and call GenerateSequence.withTimestampFn(SerializableFunction) instead
org.apache.beam.sdk.io.CountingSource.upTo(long)	use GenerateSequence instead
org.apache.beam.sdk.io.FileIO.Write.withIgnoreWindowing()	Avoid usage of this method: its effects are complex and it will be removed in future versions of Beam. Right now it exists for compatibility with WriteFiles.
org.apache.beam.sdk.io.Source.getDefaultOutputCoder()	Override Source.getOutputCoder() instead.
org.apache.beam.sdk.io.TextIO.Read.withCompressionType(TextIO.CompressionType)	Use TextIO.Read.withCompression(org.apache.beam.sdk.io.Compression).
org.apache.beam.sdk.io.TextIO.readAll()	You can achieve The functionality of TextIO.readAll() using FileIO matching plus TextIO.readFiles(). This is the preferred method to make composition explicit. TextIO.ReadAll will not receive upgrades and will be removed in a future version of Beam.
org.apache.beam.sdk.io.TextIO.ReadAll.withCompressionType(TextIO.CompressionType)	Use TextIO.ReadAll.withCompression(org.apache.beam.sdk.io.Compression).
org.apache.beam.sdk.io.TextIO.TypedWrite.to(FileBasedSink.DynamicDestinations<UserT, NewDestinationT, String>)	Use FileIO.write() or FileIO.writeDynamic() with TextIO.sink() instead.
org.apache.beam.sdk.io.TextIO.TypedWrite.withFormatFunction(SerializableFunction<UserT, String>)	Use FileIO.write() or FileIO.writeDynamic() with TextIO.sink() instead.
org.apache.beam.sdk.io.TextIO.Write.to(FileBasedSink.DynamicDestinations<String, ?, String>)	Use FileIO.write() or FileIO.writeDynamic() ()} with TextIO.sink() instead.
org.apache.beam.sdk.io.TFRecordIO.Read.withCompressionType(TFRecordIO.CompressionType)	Use TFRecordIO.Read.withCompression(org.apache.beam.sdk.io.Compression).
org.apache.beam.sdk.io.TFRecordIO.Write.withCompressionType(TFRecordIO.CompressionType)	use TFRecordIO.Write.withCompression(org.apache.beam.sdk.io.Compression).
org.apache.beam.sdk.options.ValueProviders.updateSerializedOptions(String, Map<String, String>)	Use TestPipeline.newProvider(T) for testing ValueProvider code.
org.apache.beam.sdk.Pipeline.setCoderRegistry(CoderRegistry)	this should never be used - every Pipeline has a registry throughout its lifetime.
org.apache.beam.sdk.schemas.GetterBasedSchemaProvider.fieldValueGetters(Class<?>, Schema)	new implementations should override GetterBasedSchemaProvider.fieldValueGetters(TypeDescriptor, Schema) and make this method throw an UnsupportedOperationException
org.apache.beam.sdk.schemas.GetterBasedSchemaProvider.fieldValueTypeInformations(Class<?>, Schema)	new implementations should override GetterBasedSchemaProvider.fieldValueTypeInformations(TypeDescriptor, Schema) and make this method throw an UnsupportedOperationException
org.apache.beam.sdk.schemas.GetterBasedSchemaProvider.schemaTypeCreator(Class<?>, Schema)	new implementations should override GetterBasedSchemaProvider.schemaTypeCreator(TypeDescriptor, Schema) and make this method throw an UnsupportedOperationException
org.apache.beam.sdk.schemas.parser.generated.FieldSpecifierNotationLexer.getTokenNames()
org.apache.beam.sdk.schemas.parser.generated.FieldSpecifierNotationParser.getTokenNames()
org.apache.beam.sdk.schemas.Schema.FieldType.array(Schema.FieldType, boolean)	Set the nullability on the elementType instead
org.apache.beam.sdk.schemas.Schema.FieldType.getAllMetadata()	use schema options instead.
org.apache.beam.sdk.schemas.Schema.FieldType.getMetadata(String)	use schema options instead.
org.apache.beam.sdk.schemas.Schema.FieldType.getMetadataString(String)	use schema options instead.
org.apache.beam.sdk.schemas.Schema.FieldType.map(Schema.FieldType, Schema.FieldType, boolean)	Set the nullability on the valueType instead
org.apache.beam.sdk.schemas.Schema.FieldType.withMetadata(Map<String, byte[]>)	use schema options instead.
org.apache.beam.sdk.testing.PAssert.IterableAssert.containsInAnyOrder()	Prefer PAssert.IterableAssert.empty() to this method.
org.apache.beam.sdk.testing.PAssert.PCollectionContentsAssert.equals(Object)	Object.equals(Object) is not supported on PAssert objects. If you meant to test object equality, use a variant of PAssert.PCollectionContentsAssert.containsInAnyOrder(T...) instead.
org.apache.beam.sdk.testing.PAssert.PCollectionContentsAssert.hashCode()	Object.hashCode() is not supported on PAssert objects.
org.apache.beam.sdk.transforms.DoFn.getAllowedTimestampSkew()	This method permits a DoFn to emit elements behind the watermark. These elements are considered late, and if behind the allowed lateness of a downstream PCollection may be silently dropped. See https://github.com/apache/beam/issues/18065 for details on a replacement.
org.apache.beam.sdk.transforms.DoFn.prepareForProcessing()	use DoFn.Setup or DoFn.StartBundle instead. This method will be removed in a future release.
org.apache.beam.sdk.transforms.DoFnTester.clearOutputElements()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.close()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.createProcessContext(ValueInSingleWindow<InputT>)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.finishBundle()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.getCloningBehavior()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.getMainOutputTag()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.getMutableOutput(TupleTag<T>)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.getPipelineOptions()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.of(DoFn<InputT, OutputT>)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.peekOutputElements()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.peekOutputElementsInWindow(BoundedWindow)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.peekOutputElementsWithTimestamp()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.processBundle(Iterable<? extends InputT>)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.processElement(InputT)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.processTimestampedElement(TimestampedValue<InputT>)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.processWindowedElement(InputT, Instant, BoundedWindow)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.setCloningBehavior(DoFnTester.CloningBehavior)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.setSideInput(PCollectionView<T>, BoundedWindow, T)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.setSideInputs(Map<PCollectionView<?>, Map<BoundedWindow, ?>>)	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.startBundle()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.takeOutputElements()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.DoFnTester.takeOutputElementsWithTimestamp()	Use TestPipeline with the DirectRunner.
org.apache.beam.sdk.transforms.PTransform.getDefaultOutputCoder()	Instead, the PTransform should explicitly call PCollection.setCoder(org.apache.beam.sdk.coders.Coder<T>) on the returned PCollection.
org.apache.beam.sdk.transforms.reflect.DoFnSignature.isStateful()	use DoFnSignature.usesState(), it's cleaner
org.apache.beam.sdk.transforms.View.AsMap.withSingletonValues()	this method simply returns this AsMap unmodified
org.apache.beam.sdk.transforms.View.CreatePCollectionView.getView()	This should not be used to obtain the output of any given application of this PTransform. That should be obtained by inspecting the TransformHierarchy.Node that contains this View.CreatePCollectionView, as this view may have been replaced within pipeline surgery.
org.apache.beam.sdk.transforms.windowing.WindowFn.isCompatible(WindowFn<?, ?>)	please override verifyCompatibility to throw a useful error message; we will remove isCompatible at version 3.0.0
org.apache.beam.sdk.transforms.WithTimestamps.getAllowedTimestampSkew()	This method permits a to elements to be emitted behind the watermark. These elements are considered late, and if behind the allowed lateness of a downstream PCollection may be silently dropped. See https://github.com/apache/beam/issues/18065 for details on a replacement.
org.apache.beam.sdk.transforms.WithTimestamps.withAllowedTimestampSkew(Duration)	This method permits a to elements to be emitted behind the watermark. These elements are considered late, and if behind the allowed lateness of a downstream PCollection may be silently dropped. See https://github.com/apache/beam/issues/18065 for details on a replacement.
org.apache.beam.sdk.util.construction.CreatePCollectionViewTranslation.getView(AppliedPTransform<PCollection<ElemT>, PCollection<ElemT>, PTransform<PCollection<ElemT>, PCollection<ElemT>>>)	Since View.CreatePCollectionView is not a part of the Beam model, there is no SDK-agnostic specification. Using this method means your runner is tied to Java.
org.apache.beam.sdk.values.PCollectionView.getCoderInternal()	this method will be removed entirely. The PCollection underlying a side input, including its Coder, is part of the side input's specification with a ParDo transform, which will obtain that information via a package-private channel.
org.apache.beam.sdk.values.PCollectionView.getTagInternal()	this method will be removed entirely. The PCollection underlying a side input, is part of the side input's specification with a ParDo transform, which will obtain that information via a package-private channel.
org.apache.beam.sdk.values.PCollectionView.getViewFn()	this method will be removed entirely. The ViewFn for a side input is an attribute of the side input's specification with a ParDo transform, which will obtain this specification via a package-private channel.
org.apache.beam.sdk.values.PCollectionView.getWindowingStrategyInternal()	this method will be removed entirely. The PCollection underlying a side input, including its WindowingStrategy, is part of the side input's specification with a ParDo transform, which will obtain that information via a package-private channel.
org.apache.beam.sdk.values.PCollectionViews.iterableViewUsingVoidKey(TupleTag<Materializations.MultimapView<Void, T>>, PCollection<KV<Void, T>>, PCollectionViews.TypeDescriptorSupplier<T>, WindowingStrategy<?, W>)	See PCollectionViews.iterableView(org.apache.beam.sdk.values.PCollection<T>, org.apache.beam.sdk.values.PCollectionViews.TypeDescriptorSupplier<T>, org.apache.beam.sdk.values.WindowingStrategy<?, W>).
org.apache.beam.sdk.values.PCollectionViews.listViewUsingVoidKey(TupleTag<Materializations.MultimapView<Void, T>>, PCollection<KV<Void, T>>, PCollectionViews.TypeDescriptorSupplier<T>, WindowingStrategy<?, W>)	See PCollectionViews.listView(org.apache.beam.sdk.values.PCollection<T>, org.apache.beam.sdk.values.PCollectionViews.TypeDescriptorSupplier<T>, org.apache.beam.sdk.values.WindowingStrategy<?, W>).
org.apache.beam.sdk.values.PCollectionViews.mapViewUsingVoidKey(TupleTag<Materializations.MultimapView<Void, KV<K, V>>>, PCollection<KV<Void, KV<K, V>>>, PCollectionViews.TypeDescriptorSupplier<K>, PCollectionViews.TypeDescriptorSupplier<V>, WindowingStrategy<?, W>)	See PCollectionViews.mapView(org.apache.beam.sdk.values.PCollection<org.apache.beam.sdk.values.KV<K, V>>, org.apache.beam.sdk.values.PCollectionViews.TypeDescriptorSupplier<K>, org.apache.beam.sdk.values.PCollectionViews.TypeDescriptorSupplier<V>, org.apache.beam.sdk.values.WindowingStrategy<?, W>).
org.apache.beam.sdk.values.PCollectionViews.multimapViewUsingVoidKey(TupleTag<Materializations.MultimapView<Void, KV<K, V>>>, PCollection<KV<Void, KV<K, V>>>, PCollectionViews.TypeDescriptorSupplier<K>, PCollectionViews.TypeDescriptorSupplier<V>, WindowingStrategy<?, W>)	See PCollectionViews.multimapView(org.apache.beam.sdk.values.PCollection<org.apache.beam.sdk.values.KV<K, V>>, org.apache.beam.sdk.values.PCollectionViews.TypeDescriptorSupplier<K>, org.apache.beam.sdk.values.PCollectionViews.TypeDescriptorSupplier<V>, org.apache.beam.sdk.values.WindowingStrategy<?, W>).
org.apache.beam.sdk.values.PCollectionViews.singletonViewUsingVoidKey(TupleTag<Materializations.MultimapView<Void, T>>, PCollection<KV<Void, T>>, PCollectionViews.TypeDescriptorSupplier<T>, WindowingStrategy<?, W>, boolean, T, Coder<T>)	See PCollectionViews.singletonView(org.apache.beam.sdk.values.PCollection<T>, org.apache.beam.sdk.values.PCollectionViews.TypeDescriptorSupplier<T>, org.apache.beam.sdk.values.WindowingStrategy<?, W>, boolean, T, org.apache.beam.sdk.coders.Coder<T>).
org.apache.beam.sdk.values.PValue.expand()	A PValue always expands into itself. Calling PValue.expand() on a PValue is almost never appropriate.