[
https://issues.apache.org/jira/browse/HIVE-8963?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14225455#comment-14225455
]
Jimmy Xiang commented on HIVE-8963:
-----------------------------------
Looked into it. Without limit 1, the output for spark is the same as that for
mr. With order by, the output is always right. So it should be an order isue.
Filed HIVE-8967 for trunk.
> Investigate test failure on bucketmapjoin7.q [Spark Branch]
> -----------------------------------------------------------
>
> Key: HIVE-8963
> URL: https://issues.apache.org/jira/browse/HIVE-8963
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
> Affects Versions: spark-branch
> Reporter: Chao
> Assignee: Jimmy Xiang
>
> When running on jenkins, test result for this one is different (same as MR
> though) from the result generated on my local machine. The test is similar to
> bucketmapjoin10 and bucketmapjoin11, so it might be related.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
