[ https://issues.apache.org/jira/browse/HIVE-1213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12844589#action_12844589 ]
Namit Jain commented on HIVE-1213: ---------------------------------- Consider the query; select /*+ MAPJOIN(c) */ ... from ( select /*+ MAPJOIN(b) */ .. from a join b on k1 ) subq join c on k2 Even if k1 and k2 are not same, this can be done in a singe map-only job. Note that this has nothing to do with sorting/bucketing etc. > sort-merge join does not work for sub-queries > --------------------------------------------- > > Key: HIVE-1213 > URL: https://issues.apache.org/jira/browse/HIVE-1213 > Project: Hadoop Hive > Issue Type: Bug > Components: Query Processor > Reporter: Namit Jain > Assignee: He Yongqiang > Fix For: 0.6.0 > > Attachments: Hive-1213.1.patch, Hive-1213.4.patch, Hive-1213.5.patch, > Hive-1213.6.patch, Hive-1213.7.patch > > > A query like: > select count(1) from (select /*+ MAPJOIN(x) */ from x join y ON ... ) subq; > does not work - since there is no mapping between the join operator and the > corresponding source -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.