[
https://issues.apache.org/jira/browse/TAJO-934?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14058357#comment-14058357
]
ASF GitHub Bot commented on TAJO-934:
-------------------------------------
Github user babokim commented on the pull request:
https://github.com/apache/tajo/pull/63#issuecomment-48693574
hyunsik please wait a while. I find another bug about multiple distinct.
The following query is recognized as a general group by query not a
distinct group by query.
select count(distinct col1) + count(distinct col2) from table1
I'll fix this bug in this patch.
> Multiple DISTINCT returns null grouping key value.
> --------------------------------------------------
>
> Key: TAJO-934
> URL: https://issues.apache.org/jira/browse/TAJO-934
> Project: Tajo
> Issue Type: Bug
> Reporter: Hyoungjun Kim
> Assignee: Hyoungjun Kim
> Priority: Minor
>
> {noformat}
> default> select count * from table1;
> col1, col2, col3
> -------------------------
> a, b-1, null
> a, b-2, null
> a, b-2, null
> a, b-3, null
> a, b-3, null
> a, b-3, null
> default> select col1,
> count(distinct col2) as cnt1,
> count(distinct case when col3 is not null then col2 else null end) as cnt2
> from table1
> group by col1;
> col1,cnt1,cnt2
> -------------------------------
> a, 2, 1
> null, 1, 1
> {noformat}
> This query should return the below result.
> {noformat}
> col1,cnt1,cnt2
> -------------------------------
> a, 3, 1
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.2#6252)
