[jira] [Commented] (SPARK-3198) Generates the expression id while necessary
[ https://issues.apache.org/jira/browse/SPARK-3198?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14108891#comment-14108891 ] Apache Spark commented on SPARK-3198: - User 'chenghao-intel' has created a pull request for this issue: https://github.com/apache/spark/pull/2114 > Generates the expression id while necessary > --- > > Key: SPARK-3198 > URL: https://issues.apache.org/jira/browse/SPARK-3198 > Project: Spark > Issue Type: Sub-task > Components: SQL >Reporter: Cheng Hao > > Currently, Catalyst harnesses the AtomicLong for the expression id generation > algorithm, which reduce the performance dramatically in a multithread env. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-3198) Generates the expression id while necessary
[ https://issues.apache.org/jira/browse/SPARK-3198?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14108894#comment-14108894 ] Cheng Hao commented on SPARK-3198: -- Usually, we need the expression id in logical plan analyzing, not in evaluation, hence we can get significant improvement by doing this. > Generates the expression id while necessary > --- > > Key: SPARK-3198 > URL: https://issues.apache.org/jira/browse/SPARK-3198 > Project: Spark > Issue Type: Sub-task > Components: SQL >Reporter: Cheng Hao > > Currently, Catalyst harnesses the AtomicLong for the expression id generation > algorithm, which reduce the performance dramatically in a multithread env. -- This message was sent by Atlassian JIRA (v6.2#6252) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org