[
https://issues.apache.org/jira/browse/FLINK-8689?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16394805#comment-16394805
]
ASF GitHub Bot commented on FLINK-8689:
---------------------------------------
Github user walterddr commented on the issue:
https://github.com/apache/flink/pull/5555
Thanks @hequn8128 for the prompt review.
Are you suggesting we created the mapView parallel with the accumulator?
The reason why I kept `DistinctAccumulator` is to act as a delegate to enclose
the actual accumulator so that it can be passed around in the
`accumulatorState` field without extending the arity.
I guess if we separate the mapView with the accumulator. I guess I can
separately create another field in the `accumulatorState` `Row` to store the
`mapView`(s)... This way it might be easier to handle the "reuse same mapView
for multiple different distinct agg function" case as we discussed in the doc.
Another question is I was trying to reuse as much utility of dataview
codegen as possible, as most of them are tightly coupled with the accumulators.
I guess I can further refactor (which I already did quite a bit already).
Please let me know if that's what you had in mind @hequn8128
--
Rong
> Add runtime support of distinct filter using MapView
> -----------------------------------------------------
>
> Key: FLINK-8689
> URL: https://issues.apache.org/jira/browse/FLINK-8689
> Project: Flink
> Issue Type: Sub-task
> Reporter: Rong Rong
> Assignee: Rong Rong
> Priority: Major
>
> This ticket should cover distinct aggregate function support to codegen for
> *AggregateCall*, where *isDistinct* fields is set to true.
> This can be verified using the following SQL, which is not currently
> producing correct results.
> {code:java}
> SELECT
> a,
> SUM(b) OVER (PARTITION BY a ORDER BY proctime ROWS BETWEEN 5 PRECEDING AND
> CURRENT ROW)
> FROM
> MyTable{code}
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
