[
https://issues.apache.org/jira/browse/FLINK-4692?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15817257#comment-15817257
]
ASF GitHub Bot commented on FLINK-4692:
---------------------------------------
Github user wuchong commented on the issue:
https://github.com/apache/flink/pull/2938
Thanks @twalthr for your hint. It is a bug in my
`DataSetTumbleTimeWindowAggReduceCombineFunction#combine(...)` method, that the
rowtime attribute is dropped when combining.
The collection environment will not run combine phase, but cluster
environment will. That's why we can't reproduce the wrong result in test base.
BTW, do we need to activate the cluster execution mode in table IT cases ?
Currently, only collection execution mode is activated.
> Add tumbling group-windows for batch tables
> -------------------------------------------
>
> Key: FLINK-4692
> URL: https://issues.apache.org/jira/browse/FLINK-4692
> Project: Flink
> Issue Type: Sub-task
> Components: Table API & SQL
> Reporter: Timo Walther
> Assignee: Jark Wu
>
> Add Tumble group-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
(v6.3.4#6332)
