[
https://issues.apache.org/jira/browse/SPARK-19274?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15828217#comment-15828217
]
Apache Spark commented on SPARK-19274:
--------------------------------------
User 'viirya' has created a pull request for this issue:
https://github.com/apache/spark/pull/16633
> Make GlobalLimit without shuffling data to single partition
> -----------------------------------------------------------
>
> Key: SPARK-19274
> URL: https://issues.apache.org/jira/browse/SPARK-19274
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Reporter: Liang-Chi Hsieh
>
> A logical Limit is performed actually by two physical operations LocalLimit
> and GlobalLimit.
> In most of time, before GlobalLimit, we will perform a shuffle exchange to
> shuffle data to single partition. When the limit number is not trivially
> small, this shuffling is costing.
> This change tried to perform GlobalLimit without shuffling data to single
> partition. The approach is similar to SparkPlan.executeTake. It iterates part
> of partitions until it reaches enough data.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
