Github user tianyi commented on a diff in the pull request: https://github.com/apache/spark/pull/3375#discussion_r21519366 --- Diff: sql/core/src/main/scala/org/apache/spark/sql/execution/joins/HashOuterJoin.scala --- @@ -68,62 +68,59 @@ case class HashOuterJoin( @transient private[this] lazy val DUMMY_LIST = Seq[Row](null) @transient private[this] lazy val EMPTY_LIST = Seq.empty[Row] + @transient private[this] lazy val joinedRow = new JoinedRow() --- End diff -- to @marmbrus I move `joinedRow`, `rightNullRow`, `boundCondition` to the beginning of `leftOuterIterator`, and ran my benchmark again, the result is : ``` CSV: 9935 ms CSV: 6989 ms CSV: 7045 ms Current Mem Usage:508094480 ``` without move, the result is : ``` CSV: 10590 ms CSV: 7022 ms CSV: 7086 ms Current Mem Usage:402470616 ``` (I don't know why the costs of memory is not around 208145728 in my last test, maybe something related to my laptop environment.) Apparently, the costs of memory increased 25%, and the speed is almost the same. Is it acceptable?
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org