[
https://issues.apache.org/jira/browse/HIVE-12519?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15027557#comment-15027557
]
Laljo John Pullokkaran commented on HIVE-12519:
-----------------------------------------------
You can achieve this by rewriting query with outer join & filter.
Proj(R1.*)((Filter(R2.JoinKeys is null )(R1 LOJ R2)))
There are three possible solutions:
1. User rewrites the query as above
2. Change Parser/GenPlan to do the rewrite
3. Change the Join operator to filter out tuples where RHS keys are not null.
Given the impact on rest of the optimizers, #2 may be a better option.
[~pxiong] This will be very similar to Union Distinct Rewrite.
> ANTI JOIN in hive
> -----------------
>
> Key: HIVE-12519
> URL: https://issues.apache.org/jira/browse/HIVE-12519
> Project: Hive
> Issue Type: Wish
> Reporter: Keren Edri
> Labels: hive
>
> I wish thee was "ANTI JOIN" in hive...
> please implement the ANTI JOIN as described is here
> http://blog.montmere.com/2010/12/08/the-anti-join-all-values-from-table1-where-not-in-table2/
> Thank You
> have a nice day
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
