[ https://issues.apache.org/jira/browse/SPARK-9469?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14648309#comment-14648309 ]
Apache Spark commented on SPARK-9469: ------------------------------------- User 'rxin' has created a pull request for this issue: https://github.com/apache/spark/pull/7803 > TungstenSort should not do safe -> unsafe conversion itself > ----------------------------------------------------------- > > Key: SPARK-9469 > URL: https://issues.apache.org/jira/browse/SPARK-9469 > Project: Spark > Issue Type: Sub-task > Components: SQL > Reporter: Reynold Xin > Assignee: Reynold Xin > Priority: Critical > > TungstenSort itself assumes input rows are safe rows, and uses a projection > to turn the safe rows into UnsafeRows. We should take that part of the logic > out of TungstenSort, and let the planner take care of the conversion. In that > case, if the input is UnsafeRow already, no conversion is needed. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org