[ https://issues.apache.org/jira/browse/HIVE-18008?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16244259#comment-16244259 ]
Ashutosh Chauhan commented on HIVE-18008: ----------------------------------------- {code} joinInfo.rightSet().equals(ImmutableBitSet.range(rightAggregate.getGroupCount())); {code} Just count check may not be sufficient, we should also check if they are same column. > Add optimization rule to remove gby from right side of left semi-join > --------------------------------------------------------------------- > > Key: HIVE-18008 > URL: https://issues.apache.org/jira/browse/HIVE-18008 > Project: Hive > Issue Type: Improvement > Components: Query Planning > Reporter: Vineet Garg > Assignee: Vineet Garg > Attachments: HIVE-18008.1.patch > > > Group by (on same keys as semi join) as right side of Left semi join is > unnecessary and could be removed. We see this pattern in subqueries with > explicit distinct keyword e.g. > {code:sql} > explain select * from src b where b.key in (select distinct key from src a > where a.value = b.value) > {code} -- This message was sent by Atlassian JIRA (v6.4.14#64029)