[
https://issues.apache.org/jira/browse/SPARK-11512?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15006975#comment-15006975
]
Alex Nastetsky commented on SPARK-11512:
----------------------------------------
There are 3 situations:
1) dataset A and dataset B are both partitioned/sorted the same on disk and
need to be joined. should be able to take advantage of their partitioning/sort.
2) dataset A is partitioned/sorted on disk, dataset B gets generated during the
app run and needs to be joined to dataset A. should be able to take advantage
of dataset A's partitioning/sort and mimic the same partitioning/sort on
dataset B, without having to pre-process dataset A. perhaps, something like
repartitionAndSortWithinPartitions to be performed on dataset B?
3) dataset A and B are both generated during the app run and need to be joined.
I believe doing a Sort Merge Join on these is already supported in SPARK-2213.
The first 2 situations is what this ticket is for.
> Bucket Join
> -----------
>
> Key: SPARK-11512
> URL: https://issues.apache.org/jira/browse/SPARK-11512
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Reporter: Cheng Hao
>
> Sort merge join on two datasets on the file system that have already been
> partitioned the same with the same number of partitions and sorted within
> each partition, and we don't need to sort it again while join with the
> sorted/partitioned keys
> This functionality exists in
> - Hive (hive.optimize.bucketmapjoin.sortedmerge)
> - Pig (USING 'merge')
> - MapReduce (CompositeInputFormat)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
