[
https://issues.apache.org/jira/browse/HIVE-21304?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16797647#comment-16797647
]
Zoltan Haindrich commented on HIVE-21304:
-----------------------------------------
storing the bucketingVersion in the desc made it reach places it wasn't
available before - right now I don't see the following as problematic
* cp_sel: the resultset change is due to reading 3 rows from a table; it reads
a different 3 rows
* truncate_column_buckets: the table is most probably distributed by version2 ;
so the number of rows in the 2 files have changed a little
I right now see some ways in which this bucketingVersion - or other infos may
get lost; I would rather address that in a separate ticket.
The issue is with the "clone" methods:
* abstractDesc declares that it doesn't support clones - and by doing so it
doesn't provide any facility to clone the "abstract" fields
* but there are a lot of "clone()" implementations by descendants...
* I think it might have been organized like this to try to force descendant
descs to implement it?
* instead of relying on these clone methods; I would probably go for a
kryo/unkryo to also cover all the "other" fields which we might have forgot
along the way...
> Show Bucketing version for ReduceSinkOp in explain extended plan
> ----------------------------------------------------------------
>
> Key: HIVE-21304
> URL: https://issues.apache.org/jira/browse/HIVE-21304
> Project: Hive
> Issue Type: Bug
> Reporter: Deepak Jaiswal
> Assignee: Zoltan Haindrich
> Priority: Major
> Attachments: HIVE-21304.01.patch, HIVE-21304.02.patch
>
>
> Show Bucketing version for ReduceSinkOp in explain extended plan.
> This helps identify what hashing algorithm is being used by by ReduceSinkOp.
>
> cc [~vgarg]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
