[
https://issues.apache.org/jira/browse/FLINK-15466?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Shuo Cheng updated FLINK-15466:
-------------------------------
Description:
For the sql in batch mode,
{code:java}
SELECT
a, COUNT(a), SUM(DISTINCT b) FILTER (WHERE a > 0)
FROM MyTable
GROUP BY a{code}
plan generated after logical stage is as following, which is not correct. The
`Filter $4` should be `$f2 *and* $g_0`.
!image-2020-01-03-14-24-35-943.png!
was:
For the sql in batch mode,
{code:java}
SELECT
a, COUNT(a), SUM(DISTINCT b) FILTER (WHERE a > 0)
FROM MyTable
GROUP BY a{code}
plan generated after logical stage is as following, which is not correct. The
`Filter $4` should be `$f2 *and* $g_0`.
!image-2020-01-03-14-20-54-887.png!
> `FlinkAggregateExpandDistinctAggregatesRule` generates wrong plan for cases
> that have distinct aggs with filter.
> ----------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-15466
> URL: https://issues.apache.org/jira/browse/FLINK-15466
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.8.3, 1.9.0, 1.9.1
> Reporter: Shuo Cheng
> Priority: Major
> Fix For: 1.10.0
>
> Attachments: image-2020-01-03-14-24-35-943.png
>
>
> For the sql in batch mode,
>
> {code:java}
> SELECT
> a, COUNT(a), SUM(DISTINCT b) FILTER (WHERE a > 0)
> FROM MyTable
> GROUP BY a{code}
>
> plan generated after logical stage is as following, which is not correct. The
> `Filter $4` should be `$f2 *and* $g_0`.
> !image-2020-01-03-14-24-35-943.png!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
