[
https://issues.apache.org/jira/browse/HIVE-21238?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16765269#comment-16765269
]
Vineet Garg commented on HIVE-21238:
------------------------------------
[~mgergely] Hmm now for every query irrespective of being INSERT or SELECT we
will traverse all the operators and all the column exprs within operator to get
to GenericUDFSurrogateKey. I wonder if there is a way better way to pass on
writeId to GenericUDFSurrogateKey.
> Fix Surrogate Key issue
> -----------------------
>
> Key: HIVE-21238
> URL: https://issues.apache.org/jira/browse/HIVE-21238
> Project: Hive
> Issue Type: Bug
> Components: HiveServer2
> Affects Versions: 3.0.0
> Reporter: Miklos Gergely
> Assignee: Miklos Gergely
> Priority: Major
> Fix For: 3.0.1
>
> Attachments: HIVE-21238.01.patch, HIVE-21238.02.patch
>
>
> When SemanticAnalyzer is trying to set the write id for the surrogate key
> function it is looking for such functions only in the main operator, while it
> should also look for it in it's parents.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
