[jira] [Resolved] (TRAFODION-2527) process abend when updating primary key with TRAF_RELOAD_NATABLE_CACHE 'ON'

["Suresh Subbiah (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/TRAFODION-2527?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Suresh Subbiah resolved TRAFODION-2527.
---------------------------------------
    Resolution: Fixed

> process abend when updating primary key with TRAF_RELOAD_NATABLE_CACHE 'ON'
> ---------------------------------------------------------------------------
>
>                 Key: TRAFODION-2527
>                 URL: https://issues.apache.org/jira/browse/TRAFODION-2527
>             Project: Apache Trafodion
>          Issue Type: Bug
>          Components: sql-cmp
>    Affects Versions: 2.0-incubating
>            Reporter: Suresh Subbiah
>            Assignee: Suresh Subbiah
>            Priority: Minor
>             Fix For: 2.2-incubating
>
>
> Updating primary key of a table with the cqd TRAF_RELOAD_NATABLE_CACHE 'ON' causes master executor to abend.
> To reproduce
> create table test1 like "_PRIVMGR_MD_".OBJECT_PRIVILEGES ;
> UPDATE test1  SET grantor_id = 33333, grantor_name = 'DB__ROOT',      privileges_bitmap = 3, grantable_bitmap = 0    WHERE object_uid = 6220152955843408334 AND grantor_id = 33333 AND grantee_id = 1000000 ;
> The problem is that we call getScanIndexDesc() in Inlining.cpp while transforming the tree during binding of update primary key. Now scanIndexDesc is NULL is GU natable is different from source table NATable for an update. This code is meant for triggers but the determination was incorrectly made using a pointer comparision. Now we compare using table names.



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