[
https://issues.apache.org/jira/browse/HIVE-7526?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14080630#comment-14080630
]
Xuefu Zhang commented on HIVE-7526:
-----------------------------------
Chao, thanks for your latest patch. I took the liberty of updating your patch
due to the following:
1. Your patch wasn't updated to the latest branch. Rebase was needed.
2. License header missing/removing problem.
3. More importantly, we shouldn't use a list of list to cache all rows in order
to do sortBy shuffle because of unbounded memory. We should be able to back the
returned iterator with the input iterator. I put code stubs for this.
> Research to use groupby transformation to replace Hive existing
> partitionByKey and SparkCollector combination
> -------------------------------------------------------------------------------------------------------------
>
> Key: HIVE-7526
> URL: https://issues.apache.org/jira/browse/HIVE-7526
> Project: Hive
> Issue Type: Task
> Components: Spark
> Reporter: Xuefu Zhang
> Assignee: Chao
> Attachments: HIVE-7526.2.patch, HIVE-7526.3.patch,
> HIVE-7526.4-spark.patch, HIVE-7526.5-spark.patch, HIVE-7526.patch
>
>
> Currently SparkClient shuffles data by calling paritionByKey(). This
> transformation outputs <key, value> tuples. However, Hive's ExecMapper
> expects <key, iterator<value>> tuples, and Spark's groupByKey() seems
> outputing this directly. Thus, using groupByKey, we may be able to avoid its
> own key clustering mechanism (in HiveReduceFunction). This research is to
> have a try.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
