org.apache.flink.api.java.operators

Class Grouping<T>

java.lang.Object

org.apache.flink.api.java.operators.Grouping<T>

Type Parameters:

T - The type of the elements of the grouped DataSet.

Direct Known Subclasses:

SortedGrouping, UnsortedGrouping

@Public
public abstract class Grouping<T>
extends Object

Grouping is an intermediate step for a transformation on a grouped DataSet.
The following transformation can be applied on Grouping:

• UnsortedGrouping.reduce(org.apache.flink.api.common.functions.ReduceFunction),
• UnsortedGrouping.reduceGroup(org.apache.flink.api.common.functions.GroupReduceFunction), and
• UnsortedGrouping.aggregate(org.apache.flink.api.java.aggregation.Aggregations, int).

See Also:

DataSet

Field Summary

Fields

Modifier and Type	Field and Description
protected Partitioner<?>	customPartitioner
protected DataSet<T>	inputDataSet
protected Keys<T>	keys

Constructor Summary

Constructors

Constructor and Description
Grouping(DataSet<T> set, Keys<T> keys)

Method Summary

Methods

Modifier and Type	Method and Description
Partitioner<?>	getCustomPartitioner() Gets the custom partitioner to be used for this grouping, or null, if none was defined.
DataSet<T>	getInputDataSet() Returns the input DataSet of a grouping operation, that is the one before the grouping.
Keys<T>	getKeys()

Methods inherited from class java.lang.Object

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

Field Detail

inputDataSet

protected final DataSet<T> inputDataSet

keys

protected final Keys<T> keys

customPartitioner

protected Partitioner<?> customPartitioner

Constructor Detail

Grouping

public Grouping(DataSet<T> set,
        Keys<T> keys)

Method Detail

getInputDataSet

@Internal
public DataSet<T> getInputDataSet()

Returns the input DataSet of a grouping operation, that is the one before the grouping. This means that if it is applied directly to the result of a grouping operation, it will cancel its effect. As an example, in the following snippet:

 DataSet notGrouped = input.groupBy().getDataSet();
 DataSet allReduced = notGrouped.reduce()
 

the groupBy() is as if it never happened, as the notGrouped DataSet corresponds to the input of the groupBy() (because of the getDataset()).

getKeys

@Internal
public Keys<T> getKeys()

getCustomPartitioner

@Internal
public Partitioner<?> getCustomPartitioner()

Gets the custom partitioner to be used for this grouping, or null, if none was defined.

Returns:

The custom partitioner to be used for this grouping.