[ https://issues.apache.org/jira/browse/HIVE-2238?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13572658#comment-13572658 ]
Edward Capriolo commented on HIVE-2238: --------------------------------------- We already have median, that is just percentile (50) so we may not need this. Mode is interesting. I will look over both soon. > Support for Median and Mode UDAFs > --------------------------------- > > Key: HIVE-2238 > URL: https://issues.apache.org/jira/browse/HIVE-2238 > Project: Hive > Issue Type: New Feature > Components: UDF > Reporter: Travis Powell > Labels: patch > Attachments: HIVE-2238.1.patch.txt, input.txt, median.q, > median.q.out, mode.q, mode.q.out > > > Median and Mode are essential functions for reducing/refining the data set, > and would allow for greater control over the selection of data. More involved > analytics are probably best handled by relational databases or OLAP cubes, > but Median and Mode are very practical for Hive solely in terms of delivering > a smaller data set, where items selected only have a certain mode. (Rows that > describe an object to which the table is joined where that object has a > column value frequency threshold.) > Comments are more than welcome. Would be happy to support. -- 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