[jira] [Commented] (BEAM-848) Shuffle input read-values to get maximum parallelism.
[ https://issues.apache.org/jira/browse/BEAM-848?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15953748#comment-15953748 ] Aviem Zur commented on BEAM-848: Now that BEAM-1074 is resolved a user can improve parallelism simply by setting the {{spark.default.parallelism}} configuration. [~amitsela] I believe we can close this ticket. > Shuffle input read-values to get maximum parallelism. > - > > Key: BEAM-848 > URL: https://issues.apache.org/jira/browse/BEAM-848 > Project: Beam > Issue Type: Improvement > Components: runner-spark >Reporter: Amit Sela >Assignee: Aviem Zur > > It would be wise to shuffle the read values _after_ flatmap to increase > parallelism in processing of the data. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (BEAM-848) Shuffle input read-values to get maximum parallelism.
[ https://issues.apache.org/jira/browse/BEAM-848?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15953760#comment-15953760 ] Amit Sela commented on BEAM-848: Agree. closed as invalid. If use cases prove the need, we can consider a "repartition" to max parallelism (or a pre-set one) post read. > Shuffle input read-values to get maximum parallelism. > - > > Key: BEAM-848 > URL: https://issues.apache.org/jira/browse/BEAM-848 > Project: Beam > Issue Type: Improvement > Components: runner-spark >Reporter: Amit Sela >Assignee: Aviem Zur > Fix For: First stable release > > > It would be wise to shuffle the read values _after_ flatmap to increase > parallelism in processing of the data. -- This message was sent by Atlassian JIRA (v6.3.15#6346)