[ https://issues.apache.org/jira/browse/HIVE-3562?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13546703#comment-13546703 ]
Phabricator commented on HIVE-3562: ----------------------------------- njain has commented on the revision "HIVE-3562 [jira] Some limit can be pushed down to map stage". I thought about it, even with group bys, my question is still valid. I think, there is a bug. Do you think it would be simpler to allocate a heap with (upto) topN entries instead - throw the memory threshold out. If limit < threshold, use this optimization, otherwise just ignore this optimization. INLINE COMMENTS ql/src/java/org/apache/hadoop/hive/ql/exec/ReduceSinkOperator.java:441 Isn't there a bug here ? You are using keyValues last entry to figure out whether it needs to be expanded or not. It may have an issue at the boundary - say entry 40th when a legit. entry is found. It might be simpler to pass the fact whether the entry was found or not. ql/src/java/org/apache/hadoop/hive/ql/exec/ReduceSinkOperator.java:468 This is not true if an entry is being inserted in between. I mean, if topN is 100, and we already have 100 entries. If we are inserting 50th entry, we should not be increasing usage REVISION DETAIL https://reviews.facebook.net/D5967 To: JIRA, tarball, navis Cc: njain > Some limit can be pushed down to map stage > ------------------------------------------ > > Key: HIVE-3562 > URL: https://issues.apache.org/jira/browse/HIVE-3562 > Project: Hive > Issue Type: Bug > Reporter: Navis > Assignee: Navis > Priority: Trivial > Attachments: HIVE-3562.D5967.1.patch, HIVE-3562.D5967.2.patch, > HIVE-3562.D5967.3.patch > > > Queries with limit clause (with reasonable number), for example > {noformat} > select * from src order by key limit 10; > {noformat} > makes operator tree, > TS-SEL-RS-EXT-LIMIT-FS > But LIMIT can be partially calculated in RS, reducing size of shuffling. > TS-SEL-RS(TOP-N)-EXT-LIMIT-FS -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira