org.apache.flink.table.runtime.operators.aggregate

Class MiniBatchGlobalGroupAggFunction

java.lang.Object

org.apache.flink.table.runtime.operators.bundle.MapBundleFunction<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData>

org.apache.flink.table.runtime.operators.aggregate.MiniBatchGlobalGroupAggFunction

All Implemented Interfaces:

Serializable, org.apache.flink.api.common.functions.Function

public class MiniBatchGlobalGroupAggFunction
extends MapBundleFunction<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData>

Aggregate Function used for the global groupby (without window) aggregate in miniBatch mode.

See Also:

Serialized Form

Field Summary

Fields inherited from class org.apache.flink.table.runtime.operators.bundle.MapBundleFunction

ctx

Constructor Summary

Constructors

Constructor and Description
MiniBatchGlobalGroupAggFunction(GeneratedAggsHandleFunction genLocalAggsHandler, GeneratedAggsHandleFunction genGlobalAggsHandler, GeneratedRecordEqualiser genRecordEqualiser, org.apache.flink.table.types.logical.LogicalType[] accTypes, int indexOfCountStar, boolean generateUpdateBefore, long stateRetentionTime) Creates a MiniBatchGlobalGroupAggFunction.

Method Summary

Modifier and Type	Method and Description
org.apache.flink.table.data.RowData	addInput(org.apache.flink.table.data.RowData previousAcc, org.apache.flink.table.data.RowData input) The previousAcc is accumulator, but input is a row in <key, accumulator> schema, the specific generated localAgg will project the input to accumulator in merge method.
void	close()
void	finishBundle(Map<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData> buffer, org.apache.flink.util.Collector<org.apache.flink.table.data.RowData> out) Called when a bundle is finished.
void	open(ExecutionContext ctx)

Methods inherited from class java.lang.Object

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

Constructor Detail

MiniBatchGlobalGroupAggFunction

public MiniBatchGlobalGroupAggFunction(GeneratedAggsHandleFunction genLocalAggsHandler,
                                       GeneratedAggsHandleFunction genGlobalAggsHandler,
                                       GeneratedRecordEqualiser genRecordEqualiser,
                                       org.apache.flink.table.types.logical.LogicalType[] accTypes,
                                       int indexOfCountStar,
                                       boolean generateUpdateBefore,
                                       long stateRetentionTime)

Creates a MiniBatchGlobalGroupAggFunction.

Parameters:

genLocalAggsHandler - The generated local aggregate handler

genGlobalAggsHandler - The generated global aggregate handler

genRecordEqualiser - The code generated equaliser used to equal RowData.

accTypes - The accumulator types.

indexOfCountStar - The index of COUNT(*) in the aggregates. -1 when the input doesn't contain COUNT(*), i.e. doesn't contain UPDATE_BEFORE or DELETE messages. We make sure there is a COUNT(*) if input stream contains UPDATE_BEFORE or DELETE messages.

generateUpdateBefore - Whether this operator will generate UPDATE_BEFORE messages.

stateRetentionTime - state idle retention time which unit is MILLISECONDS.

Method Detail

open

public void open(ExecutionContext ctx)
          throws Exception

Overrides:

open in class MapBundleFunction<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData>

Throws:

Exception

addInput

public org.apache.flink.table.data.RowData addInput(@Nullable
                                                    org.apache.flink.table.data.RowData previousAcc,
                                                    org.apache.flink.table.data.RowData input)
                                             throws Exception

The previousAcc is accumulator, but input is a row in <key, accumulator> schema, the specific generated localAgg will project the input to accumulator in merge method.

Specified by:

addInput in class MapBundleFunction<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData>

Parameters:

previousAcc - the existing bundle value, maybe null

input - the given input, not null

Throws:

Exception

finishBundle

public void finishBundle(Map<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData> buffer,
                         org.apache.flink.util.Collector<org.apache.flink.table.data.RowData> out)
                  throws Exception

Description copied from class: MapBundleFunction

Called when a bundle is finished. Transform a bundle to zero, one, or more output elements.

Specified by:

finishBundle in class MapBundleFunction<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData>

Throws:

Exception

close

public void close()
           throws Exception

Overrides:

close in class MapBundleFunction<org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData,org.apache.flink.table.data.RowData>

Throws:

Exception