[ https://issues.apache.org/jira/browse/HIVE-8667?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14192167#comment-14192167 ]
Ashutosh Chauhan commented on HIVE-8667: ---------------------------------------- It seems difference here stems of what rules Hive follows for PPD vs what are followed in Calcite. Hive PPD pushes filter through both branches of join to table scans, whereas Calcite is pushing filter only to outer table. Later in compilation, NullScanOptimizer found that filter only on side and converted to null-scan, but didnt filter on other side, so didnt do anything with it. Also, Calcite changed left outer join to right outer join and swapped tables on both sides, along with filter. > CBO: optimize_nullscan - some TableScans for nullscans appeared after CBO > ------------------------------------------------------------------------- > > Key: HIVE-8667 > URL: https://issues.apache.org/jira/browse/HIVE-8667 > Project: Hive > Issue Type: Sub-task > Components: CBO > Reporter: Sergey Shelukhin > Assignee: Sergey Shelukhin > Fix For: 0.15.0 > > > Looks like some rewriting by CBO prevents nullscans from being optimized. -- This message was sent by Atlassian JIRA (v6.3.4#6332)