Package org.apache.flink.table.connector.source.abilities

Interface SupportsLookupCustomShuffle

@PublicEvolving
public interface SupportsLookupCustomShuffle

This interface is designed to allow connectors to provide a custom partitioning strategy for the data that is fed into the LookupTableSource. This enables the Flink Planner to optimize the distribution of input stream across different subtasks of lookup-join node to match the distribution of data in the external data source.

Nested Class Summary

Nested Classes

Modifier and Type	Interface	Description
static interface	SupportsLookupCustomShuffle.InputDataPartitioner	This interface is responsible for providing custom partitioning logic for the RowData records.

Method Summary

Modifier and Type	Method	Description
Optional<SupportsLookupCustomShuffle.InputDataPartitioner>	getPartitioner()	This method is used to retrieve a custom partitioner that will be applied to the input stream of lookup-join node.

Method Details

getPartitioner

Optional<SupportsLookupCustomShuffle.InputDataPartitioner> getPartitioner()

This method is used to retrieve a custom partitioner that will be applied to the input stream of lookup-join node.

Returns:

An SupportsLookupCustomShuffle.InputDataPartitioner that defines how records should be distributed across the different subtasks. If the connector expects the input data to remain in its original distribution, an Optional.empty() should be returned.