[
https://issues.apache.org/jira/browse/PHOENIX-1168?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14145025#comment-14145025
]
Maryann Xue commented on PHOENIX-1168:
--------------------------------------
[~jamestaylor] Good point. I've been thinking about this too. Actually
non-correlated subqueries and EXISTS/NOT EXISTS correlated subqueries are most
of the time equivalent. So let me take that (abstraction for different code
paths) into account when I work on many-to-many joins. The client-side
many-to-many joins would have one limitation compared to hash join though, for
it can only support two joining tables and one-level sub-queries without
introducing temporary tables.
So how about I check-in this first and open another two entries for (1)
different code path for sub-query execution (2) ANY/SOME/ALL rewrite ?
> Support non-correlated sub-queries in where clause having a comparison
> operator with no modifier or a comparison operator modified by ANY, SOME or
> ALL
> ------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: PHOENIX-1168
> URL: https://issues.apache.org/jira/browse/PHOENIX-1168
> Project: Phoenix
> Issue Type: Sub-task
> Affects Versions: 3.0.0, 4.0.0, 5.0.0
> Reporter: Maryann Xue
> Assignee: Maryann Xue
> Fix For: 3.0.0, 4.0.0, 5.0.0
>
> Attachments: 1168.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
