[
https://issues.apache.org/jira/browse/HIVE-12590?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15051600#comment-15051600
]
Laljo John Pullokkaran commented on HIVE-12590:
-----------------------------------------------
Obvious change was to remove lowercase conversion in Phase1.
However it seems deeper than that.
[~ashutoshc] Could you take a look? I have other stuff in my plate right now.
> Repeated UDAFs with literals can produce incorrect result
> ---------------------------------------------------------
>
> Key: HIVE-12590
> URL: https://issues.apache.org/jira/browse/HIVE-12590
> Project: Hive
> Issue Type: Bug
> Components: Query Planning
> Affects Versions: 1.0.1, 1.1.1, 1.2.1, 2.0.0
> Reporter: Laljo John Pullokkaran
> Assignee: Laljo John Pullokkaran
> Priority: Critical
>
> Repeated UDAF with literals could produce wrong result.
> This is not a common use case, nevertheless a bug.
> hive> select max('pants'), max('pANTS') from t1 group by key;
> Total MapReduce CPU Time Spent: 0 msec
> OK
> pANTS pANTS
> pANTS pANTS
> pANTS pANTS
> pANTS pANTS
> pANTS pANTS
> Time taken: 296.252 seconds, Fetched: 5 row(s)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
