[
https://issues.apache.org/jira/browse/HIVE-8745?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14199970#comment-14199970
]
Gunther Hagleitner commented on HIVE-8745:
------------------------------------------
[~xuefuz] [~jdere] is right. You can't have it both ways. I don't see how you
create an object that compares as equal on the byte-level but then magically
reconstructs additional information on deserialization. You could add info to
the value part of the MR key/value tuple but that's an unnecessarily complex
solution. As [~jdere] says: This is a regression and I think we should revert
HIVE-7373.
The other option would be to pad all values to the column spec and make sure we
compute the spec as the max for the join keys. I'm not sure why you were
against that in the first place - it seems that's what most DBs do. However,
that's complicated and should be tackled in 0.15.0.
> Joins on decimal keys return different results whether they are run as reduce
> join or map join
> ----------------------------------------------------------------------------------------------
>
> Key: HIVE-8745
> URL: https://issues.apache.org/jira/browse/HIVE-8745
> Project: Hive
> Issue Type: Bug
> Affects Versions: 0.14.0
> Reporter: Gunther Hagleitner
> Assignee: Jason Dere
> Priority: Critical
> Fix For: 0.14.0
>
> Attachments: join_test.q
>
>
> See attached .q file to reproduce. The difference seems to be whether
> trailing 0s are considered the same value or not.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
