Package org.apache.iceberg.flink.sink

Class PartitionKeySelector

java.lang.Object

org.apache.iceberg.flink.sink.PartitionKeySelector

All Implemented Interfaces:

Serializable, org.apache.flink.api.common.functions.Function, org.apache.flink.api.java.functions.KeySelector<org.apache.flink.table.data.RowData,String>

@Internal
public class PartitionKeySelector
extends Object
implements org.apache.flink.api.java.functions.KeySelector<org.apache.flink.table.data.RowData,String>

Create a KeySelector to shuffle by partition key, then each partition/bucket will be wrote by only one task. That will reduce lots of small files in partitioned fanout write policy for FlinkSink.

See Also:

• Serialized Form

Constructor Summary

Constructors

Constructor	Description
PartitionKeySelector(PartitionSpec spec, Schema schema, org.apache.flink.table.types.logical.RowType flinkSchema)

Method Summary

Modifier and Type	Method	Description
String	getKey(org.apache.flink.table.data.RowData row)

Methods inherited from class java.lang.Object

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

Constructor Details

PartitionKeySelector

public PartitionKeySelector(PartitionSpec spec,
 Schema schema,
 org.apache.flink.table.types.logical.RowType flinkSchema)

Method Details

getKey

public String getKey(org.apache.flink.table.data.RowData row)

Specified by:

getKey in interface org.apache.flink.api.java.functions.KeySelector<org.apache.flink.table.data.RowData,String>