Combine.PerKey (Google Cloud Dataflow SDK 1.9.1 API)

Google Cloud Dataflow SDK for Java, version 1.9.1

Warning: Dataflow SDK 1.x for Java is unsupported as of October 16, 2018. See Migrating from Dataflow SDK 1.x for Java for migration guidance.

Note: If you are using Cloud Dataflow SDK version 2.x, see the Apache Beam API reference documentation.

com.google.cloud.dataflow.sdk.transforms

Class Combine.PerKey<K,InputT,OutputT>

• java.lang.Object
• • com.google.cloud.dataflow.sdk.transforms.PTransform<PCollection<KV<K,InputT>>,PCollection<KV<K,OutputT>>>
  • • com.google.cloud.dataflow.sdk.transforms.Combine.PerKey<K,InputT,OutputT>

• Type Parameters:

  K - the type of the keys of the input and output PCollections

  InputT - the type of the values of the input PCollection

  OutputT - the type of the values of the output PCollection

  All Implemented Interfaces:

  HasDisplayData, Serializable

  Enclosing class:

  Combine

  ---

  
  

  public static class Combine.PerKey<K,InputT,OutputT>
  extends PTransform<PCollection<KV<K,InputT>>,PCollection<KV<K,OutputT>>>

  PerKey<K, InputT, OutputT> takes a PCollection<KV<K, InputT>>, groups it by key, applies a combining function to the InputT values associated with each key to produce a combined OutputT value, and returns a PCollection<KV<K, OutputT>> representing a map from each distinct key of the input PCollection to the corresponding combined value. InputT and OutputT are often the same.

  This is a concise shorthand for an application of GroupByKey followed by an application of Combine.GroupedValues. See those operations for more details on how keys are compared for equality and on the default Coder for the output.

  Example of use:

   
   PCollection<KV<String, Double>> salesRecords = ...;
   PCollection<KV<String, Double>> totalSalesPerPerson =
       salesRecords.apply(Combine.<String, Double>perKey(
           new Sum.SumDoubleFn()));
    

  Each output element is in the window by which its corresponding input was grouped, and has the timestamp of the end of that window. The output PCollection has the same WindowFn as the input.

  See Also:

  Serialized Form

• • Field Summary

    • Fields inherited from class com.google.cloud.dataflow.sdk.transforms.PTransform

      name

  • Method Summary

    All Methods Instance Methods Concrete Methods 

    Modifier and Type	Method and Description
    PCollection<KV<K,OutputT>>	apply(PCollection<KV<K,InputT>> input) Applies this PTransform on the given InputT, and returns its Output.
    CombineFnBase.PerKeyCombineFn<? super K,? super InputT,?,OutputT>	getFn() Returns the CombineFnBase.PerKeyCombineFn used by this Combine operation.
    List<PCollectionView<?>>	getSideInputs() Returns the side inputs used by this Combine operation.
    Combine.PerKey<K,InputT,OutputT>	named(String name) Return a new Globally transform that's like this transform but with the specified name.
    void	populateDisplayData(DisplayData.Builder builder) Register display data for the given transform or component.
    Combine.PerKeyWithHotKeyFanout<K,InputT,OutputT>	withHotKeyFanout(int hotKeyFanout) Like withHotKeyFanout(SerializableFunction), but returning the given constant value for every key.
    Combine.PerKeyWithHotKeyFanout<K,InputT,OutputT>	withHotKeyFanout(SerializableFunction<? super K,Integer> hotKeyFanout) If a single key has disproportionately many values, it may become a bottleneck, especially in streaming mode.
    Combine.PerKey<K,InputT,OutputT>	withSideInputs(Iterable<? extends PCollectionView<?>> sideInputs) Returns a PTransform identical to this, but with the specified side inputs to use in CombineWithContext.KeyedCombineFnWithContext.

    • Methods inherited from class com.google.cloud.dataflow.sdk.transforms.PTransform

      getDefaultOutputCoder, getDefaultOutputCoder, getDefaultOutputCoder, getKindString, getName, toString, validate

    • Methods inherited from class java.lang.Object

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

• • Method Detail

    • named

      public Combine.PerKey<K,InputT,OutputT> named(String name)

      Return a new Globally transform that's like this transform but with the specified name. Does not modify this transform.

    • withSideInputs

      public Combine.PerKey<K,InputT,OutputT> withSideInputs(Iterable<? extends PCollectionView<?>> sideInputs)

      Returns a PTransform identical to this, but with the specified side inputs to use in CombineWithContext.KeyedCombineFnWithContext.

    • withHotKeyFanout

      public Combine.PerKeyWithHotKeyFanout<K,InputT,OutputT> withHotKeyFanout(SerializableFunction<? super K,Integer> hotKeyFanout)

      If a single key has disproportionately many values, it may become a bottleneck, especially in streaming mode. This returns a new per-key combining transform that inserts an intermediate node to combine "hot" keys partially before performing the full combine.

      Parameters:

      hotKeyFanout - a function from keys to an integer N, where the key will be spread among N intermediate nodes for partial combining. If N is less than or equal to 1, this key will not be sent through an intermediate node.

    • withHotKeyFanout

      public Combine.PerKeyWithHotKeyFanout<K,InputT,OutputT> withHotKeyFanout(int hotKeyFanout)

      Like withHotKeyFanout(SerializableFunction), but returning the given constant value for every key.

    • getFn

      public CombineFnBase.PerKeyCombineFn<? super K,? super InputT,?,OutputT> getFn()

      Returns the CombineFnBase.PerKeyCombineFn used by this Combine operation.

    • getSideInputs

      public List<PCollectionView<?>> getSideInputs()

      Returns the side inputs used by this Combine operation.

    • apply

      public PCollection<KV<K,OutputT>> apply(PCollection<KV<K,InputT>> input)

      Description copied from class: PTransform
      Applies this PTransform on the given InputT, and returns its Output.

      Composite transforms, which are defined in terms of other transforms, should return the output of one of the composed transforms. Non-composite transforms, which do not apply any transforms internally, should return a new unbound output and register evaluators (via backend-specific registration methods).

      The default implementation throws an exception. A derived class must either implement apply, or else each runner must supply a custom implementation via PipelineRunner.apply(com.google.cloud.dataflow.sdk.transforms.PTransform<InputT, OutputT>, InputT).

      Overrides:

      apply in class PTransform<PCollection<KV<K,InputT>>,PCollection<KV<K,OutputT>>>

    • populateDisplayData

      public void populateDisplayData(DisplayData.Builder builder)

      Description copied from class: PTransform
      Register display data for the given transform or component.

      populateDisplayData(DisplayData.Builder) is invoked by Pipeline runners to collect display data via DisplayData.from(HasDisplayData). Implementations may call super.populateDisplayData(builder) in order to register display data in the current namespace, but should otherwise use subcomponent.populateDisplayData(builder) to use the namespace of the subcomponent.

      By default, does not register any display data. Implementors may override this method to provide their own display data.

      Specified by:

      populateDisplayData in interface HasDisplayData

      Overrides:

      populateDisplayData in class PTransform<PCollection<KV<K,InputT>>,PCollection<KV<K,OutputT>>>

      Parameters:

      builder - The builder to populate with display data.

      See Also:

      HasDisplayData