[ 
https://issues.apache.org/jira/browse/HIVE-8151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14151227#comment-14151227
 ] 

Prasanth J commented on HIVE-8151:
----------------------------------

[~wzc1989] There are 2 issues with this optimization which was found recently. 
One is this issue (HIVE-8151) where the last record of a particular group by 
ends up in the next partition. Other one is HIVE-8162 where in case of group-by 
additional columns are added to reducer key which should have been done only 
for order-by queries. Does any of these issues solve your case? I am still not 
sure how ClassCastException is related to any of these cases. Would it be 
possible for you check if HIVE-8151 or HIVE-8162 solves your issue? 
Alternatively, you can provide me a small reproducible case which I can use to 
make sure this feature works properly.

> Dynamic partition sort optimization inserts record wrongly to partition when 
> used with GroupBy
> ----------------------------------------------------------------------------------------------
>
>                 Key: HIVE-8151
>                 URL: https://issues.apache.org/jira/browse/HIVE-8151
>             Project: Hive
>          Issue Type: Bug
>    Affects Versions: 0.14.0, 0.13.1
>            Reporter: Prasanth J
>            Assignee: Prasanth J
>            Priority: Blocker
>         Attachments: HIVE-8151.1.patch, HIVE-8151.2.patch, HIVE-8151.3.patch
>
>
> HIVE-6455 added dynamic partition sort optimization. It added startGroup() 
> method to FileSink operator to look for changes in reduce key for creating 
> partition directories. This method however is not reliable as the key called 
> with startGroup() is different from the key called with processOp(). 
> startGroup() is called with newly changed key whereas processOp() is called 
> with previously aggregated key. This will result in processOp() writing the 
> last row of previous group as the first row of next group. This happens only 
> when used with group by operator.
> The fix is to not rely on startGroup() and do the partition directory 
> creation in processOp() itself.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to