Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3585#discussion_r107866801
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamOverAggregate.scala
---
@@ -130,32 +169,76 @@ class DataStreamOverAggregate(
val rowTypeInfo =
FlinkTypeFactory.toInternalRowTypeInfo(getRowType).asInstanceOf[RowTypeInfo]
val result: DataStream[Row] =
- // partitioned aggregation
- if (partitionKeys.nonEmpty) {
- val processFunction =
AggregateUtil.CreateUnboundedProcessingOverProcessFunction(
- namedAggregates,
- inputType)
+ // partitioned aggregation
+ if (partitionKeys.nonEmpty) {
+ val processFunction =
AggregateUtil.createUnboundedProcessingOverProcessFunction(
+ namedAggregates,
+ inputType)
- inputDS
+ inputDS
.keyBy(partitionKeys: _*)
.process(processFunction)
.returns(rowTypeInfo)
.name(aggOpName)
.asInstanceOf[DataStream[Row]]
- }
- // non-partitioned aggregation
- else {
- val processFunction =
AggregateUtil.CreateUnboundedProcessingOverProcessFunction(
- namedAggregates,
- inputType,
- false)
-
- inputDS
-
.process(processFunction).setParallelism(1).setMaxParallelism(1)
+ }
+ // non-partitioned aggregation
+ else {
+ val processFunction =
AggregateUtil.createUnboundedProcessingOverProcessFunction(
+ namedAggregates,
+ inputType,
+ false)
+
+ inputDS
+ .process(processFunction).setParallelism(1).setMaxParallelism(1)
+ .returns(rowTypeInfo)
+ .name(aggOpName)
+ .asInstanceOf[DataStream[Row]]
+ }
+ result
+ }
+
+ def createRowsClauseBoundedAndCurrentRowOverWindow(
+ inputDS: DataStream[Row],
+ isRowTimeType: Boolean = false): DataStream[Row] = {
+
+ val overWindow: Group = logicWindow.groups.get(0)
+ val partitionKeys: Array[Int] = overWindow.keys.toArray
+ val namedAggregates: Seq[CalcitePair[AggregateCall, String]] =
generateNamedAggregates
+ val inputFields = (0 until inputType.getFieldCount).toArray
+
+ val precedingOffset =
+ getLowerBoundary(logicWindow, overWindow, getInput()) + 1
+
+ // get the output types
+ val rowTypeInfo =
FlinkTypeFactory.toInternalRowTypeInfo(getRowType).asInstanceOf[RowTypeInfo]
+
+ val processFunction =
AggregateUtil.createRowsClauseBoundedOverProcessFunction(
+ namedAggregates,
+ inputType,
+ inputFields,
+ precedingOffset,
+ isRowTimeType
+ )
+ val result: DataStream[Row] =
+ // partitioned aggregation
+ if (partitionKeys.nonEmpty) {
+ inputDS
+ .keyBy(partitionKeys: _*)
+ .process(processFunction)
.returns(rowTypeInfo)
.name(aggOpName)
.asInstanceOf[DataStream[Row]]
- }
+ }
+ // non-partitioned aggregation
+ else {
+ inputDS
+ .keyBy(new NullByteKeySelector[Row])
+ .process(processFunction)
--- End diff --
`setParallelism()` and `setMaxParallelism()`
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
