[
https://issues.apache.org/jira/browse/FLINK-29692?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17622940#comment-17622940
]
Jing Zhang edited comment on FLINK-29692 at 10/24/22 4:21 AM:
--------------------------------------------------------------
Sorry for late response.
I'm still wondering whether we really need early fire/late fire.
As [~jark] said, eary fire/late fires is supported in group window aggregation.
when users enable early fire/ late fires of window aggregate, he/she would
allow:
# the emitted result would has no restrict mapping with time
# the intermediate result could be allowed to emitted
# the window result might be retracted later
If a user could allows the above conditions, why not use unbounded aggregate
directly? Otherwise, window aggregate would be the best options.
Could you describe the demand in detail [~canopenerda] [~jjpar] , thanks a lot.
Let's see if there are other better plans which could satisfy your demands,
such as [cumulate window
tvf?|https://nightlies.apache.org/flink/flink-docs-release-1.14/zh/docs/dev/table/sql/queries/window-tvf/#cumulate]
was (Author: qingru zhang):
Sorry for late response.
I'm still wondering whether we really need early fire/late fire.
As [~jark] said, eary fire/late fires is supported in group window aggregation.
when users enable early fire/ late fires of window aggregate, he/she would
allow:
# the emitted result would has no restrict mapping with time
# the intermediate result could be allowed to emitted
# the window result might be retracted later
If a user could allows the above conditions, why not use unbounded aggregate
directly? Otherwise, window aggregate would be the best options.
Could you describe the demand in detail [~canopenerda] [~jjpar] , thanks a lot.
Maybe there are others plans which could satisfy your demands. such as
[cumulate window
tvf|https://nightlies.apache.org/flink/flink-docs-release-1.14/zh/docs/dev/table/sql/queries/window-tvf/#cumulate]
> Support early/late fires for Windowing TVFs
> -------------------------------------------
>
> Key: FLINK-29692
> URL: https://issues.apache.org/jira/browse/FLINK-29692
> Project: Flink
> Issue Type: New Feature
> Components: Table SQL / Planner
> Affects Versions: 1.15.2
> Reporter: Canope Nerda
> Priority: Major
>
> I have cases where I need to 1) output data as soon as possible and 2) handle
> late arriving data to achieve eventual correctness. In the logic, I need to
> do window deduplication which is based on Windowing TVFs and according to
> source code, early/late fires are not supported yet in Windowing TVFs.
> Actually 1) contradicts with 2). Without early/late fires, we had to
> compromise, either live with fresh incorrect data or tolerate excess latency
> for correctness.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
