[
https://issues.apache.org/jira/browse/FLINK-22781?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17353126#comment-17353126
]
Shuo Cheng commented on FLINK-22781:
------------------------------------
[~qingru zhang], I think what you mentioned above is the main concern why
window do not support updating input stream before, either for CDC source or
updating stream from other operator (e.g, rowtime deduplicate), because we do
not know when the updating message for a specific key will happen, and window
state will be cleaned once the window is fired. It is possible to keep the
result correct by increasing the ttl for window (e.g, set allow lateness for
window), but I'think that's not an elegant solution...
> Incorrect result for group window aggregate when mini-batch is enabled
> ----------------------------------------------------------------------
>
> Key: FLINK-22781
> URL: https://issues.apache.org/jira/browse/FLINK-22781
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.14.0
> Reporter: godfrey he
> Assignee: Andy
> Priority: Critical
> Fix For: 1.14.0
>
>
> We can reproduce this issue through adding the following code to
> {{GroupWindowITCase#testWindowAggregateOnUpsertSource}} method:
> {code:java}
> tEnv.getConfig.getConfiguration.setBoolean(
> ExecutionConfigOptions.TABLE_EXEC_MINIBATCH_ENABLED, true)
> tEnv.getConfig.getConfiguration.set(
> ExecutionConfigOptions.TABLE_EXEC_MINIBATCH_ALLOW_LATENCY,
> Duration.ofSeconds(1))
> tEnv.getConfig.getConfiguration.setLong(
> ExecutionConfigOptions.TABLE_EXEC_MINIBATCH_SIZE, 10L)
> {code}
> The reason is the group window without any data (the data may be retracted)
> should not send any record.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
