You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Szehon Ho (JIRA)" <ji...@apache.org> on 2015/09/28 21:13:04 UTC
[jira] [Commented] (HIVE-11962) Improve windowing_windowspec2.q
tests to return consistent results
[ https://issues.apache.org/jira/browse/HIVE-11962?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14933827#comment-14933827 ]
Szehon Ho commented on HIVE-11962:
----------------------------------
Makes sense to me, thanks Aihua +1
> Improve windowing_windowspec2.q tests to return consistent results
> ------------------------------------------------------------------
>
> Key: HIVE-11962
> URL: https://issues.apache.org/jira/browse/HIVE-11962
> Project: Hive
> Issue Type: Test
> Components: Test
> Affects Versions: 2.0.0
> Reporter: Aihua Xu
> Assignee: Aihua Xu
> Priority: Trivial
> Attachments: HIVE-11962.patch
>
>
> Upstream test result for windowing_windowspec2.q seems to return consistent result while we have observe that in a different test env, the result could be slightly different.
> e.g., for the following query, the value t could be the same in each partition of ts. So the row order could be either for those rows. I haven't looked further why it causes that difference yet.
> {noformat}
> select ts, f, max(f) over (partition by ts order by t rows between 2 preceding and 1 preceding) from over10k limit 100;
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)