org.apache.hudi.table.action.cluster.strategy

Class UpdateStrategy<T,I>

java.lang.Object

org.apache.hudi.table.action.cluster.strategy.UpdateStrategy<T,I>

All Implemented Interfaces:

Serializable

public abstract class UpdateStrategy<T,I>
extends Object
implements Serializable

When file groups in clustering, write records to these file group need to check.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected HoodieEngineContext	engineContext
protected Set<HoodieFileGroupId>	fileGroupsInPendingClustering
protected HoodieTable	table

Constructor Summary

Constructors

Constructor and Description
UpdateStrategy(HoodieEngineContext engineContext, HoodieTable table, Set<HoodieFileGroupId> fileGroupsInPendingClustering)

Method Summary

Modifier and Type	Method and Description
abstract Pair<I,Set<HoodieFileGroupId>>	handleUpdate(I taggedRecordsRDD) Check the update records to the file group in clustering.

Methods inherited from class java.lang.Object

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

Field Detail

engineContext

protected final transient HoodieEngineContext engineContext

table

protected HoodieTable table

fileGroupsInPendingClustering

protected Set<HoodieFileGroupId> fileGroupsInPendingClustering

Constructor Detail

UpdateStrategy

public UpdateStrategy(HoodieEngineContext engineContext,
                      HoodieTable table,
                      Set<HoodieFileGroupId> fileGroupsInPendingClustering)

Method Detail

handleUpdate

public abstract Pair<I,Set<HoodieFileGroupId>> handleUpdate(I taggedRecordsRDD)

Check the update records to the file group in clustering.

Parameters:

taggedRecordsRDD - the records to write, tagged with target file id, future can update tagged records location to a different fileId.

Returns:

the recordsRDD strategy updated and a set of file groups to be updated while pending clustering.