[teiid-issues] [JBoss JIRA] (TEIID-3511) usage of BITAND function in exists statement results in duplicate rows

[Steven Hawkins (JIRA)]

    [ https://issues.jboss.org/browse/TEIID-3511?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13081696#comment-13081696 ] 

Steven Hawkins commented on TEIID-3511:
---------------------------------------

>  I'm sorry but I can not give you the data because it is medical data. I tried to reproduce this with less joins but then everything seems to work out fine.

Would it be possible to provide a post execution query plan that shows the row counts for the processing nodes?  That may provide a clue as to where the processing is not correct.

> Do you have any idea what might have caused this?

Bitand is not a source supported function (which sources are you accessing?) and so the subquery does not pushdown and the resulting plan is different than without the bitand predicate.  From there it's not clear where the problem is.

> I now also used a temp table joined with 2 tables and then a where clause that has an exists condition with and without a BITAND function call.

Are you saying that it also appears to return duplicate results?

> usage of BITAND function in exists statement results in duplicate rows
> ----------------------------------------------------------------------
>
>                 Key: TEIID-3511
>                 URL: https://issues.jboss.org/browse/TEIID-3511
>             Project: Teiid
>          Issue Type: Bug
>    Affects Versions: 8.9.1
>            Reporter: Bram Gadeyne
>            Assignee: Steven Hawkins
>             Fix For: 8.12
>
>         Attachments: debugplan_bitand.txt, debugplan_bitand_correct.txt, debugplan_nobitand.txt, debugplan_nobitand_correct.txt
>
>
> I've added the debug plan for the query without the BITAND function and with the BITAND function as an attachment.
> The version without the BITAND function returns 438 rows and the version with BITAND function returns 833 rows. I can see however that rows are duplicated in this second result. I've checked this by executing the version with BITAND statement with select *.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)