[
https://issues.apache.org/jira/browse/FLINK-5219?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15865639#comment-15865639
]
ASF GitHub Bot commented on FLINK-5219:
---------------------------------------
Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3266#discussion_r101015171
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/aggregate/DataSetSessionWindowAggregateProcessor.scala
---
@@ -59,7 +61,9 @@ class DataSetSessionWindowAggregateReduceGroupFunction(
finalRowWindowEndPos: Option[Int],
gap:Long,
isInputCombined: Boolean)
- extends RichGroupReduceFunction[Row, Row] {
+ extends AbstractRichFunction
+ with MapPartitionFunction[Row, Row]
--- End diff --
I like the idea of implementing a joint `MapPartition` and `GroupReduce`
function. However, I think it is not necessary for the final aggregation. We
can also call `DataSet.reduceGroup()` which will do the same as
`DataSet.mapPartition().parallelism(1)`.
> Add non-grouped session windows for batch tables
> ------------------------------------------------
>
> Key: FLINK-5219
> URL: https://issues.apache.org/jira/browse/FLINK-5219
> Project: Flink
> Issue Type: Sub-task
> Components: Table API & SQL
> Reporter: sunjincheng
> Assignee: sunjincheng
>
> Add non-grouped session windows for batch tables as described in
> [FLIP-11|https://cwiki.apache.org/confluence/display/FLINK/FLIP-11%3A+Table+API+Stream+Aggregations].
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
