[
https://issues.apache.org/jira/browse/SPARK-24528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16533339#comment-16533339
]
Ohad Raviv commented on SPARK-24528:
------------------------------------
After digging a little bit in the code and Jira I understand that this is just
a special case of SPARK-2926, just that the performance boost is greater.
over there they deal with moving the sort work from reducers to mappers and
show reducers performance boost of ~10x and an overall performance boost of ~2x
(I'm not sure why it has never got merged). In our case because the data is
already sorted in the buckets we should expect this great 10x boost!
because most of the needed code is already in there I guess it will be wise to
migrate it (altough it contains some more fancy things like Tiered Merger that
I'm not sure we need).
> Missing optimization for Aggregations/Windowing on a bucketed table
> -------------------------------------------------------------------
>
> Key: SPARK-24528
> URL: https://issues.apache.org/jira/browse/SPARK-24528
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 2.3.0, 2.4.0
> Reporter: Ohad Raviv
> Priority: Major
>
> Closely related to SPARK-24410, we're trying to optimize a very common use
> case we have of getting the most updated row by id from a fact table.
> We're saving the table bucketed to skip the shuffle stage, but we're still
> "waste" time on the Sort operator evethough the data is already sorted.
> here's a good example:
> {code:java}
> sparkSession.range(N).selectExpr(
> "id as key",
> "id % 2 as t1",
> "id % 3 as t2")
> .repartition(col("key"))
> .write
> .mode(SaveMode.Overwrite)
> .bucketBy(3, "key")
> .sortBy("key", "t1")
> .saveAsTable("a1"){code}
> {code:java}
> sparkSession.sql("select max(struct(t1, *)) from a1 group by key").explain
> == Physical Plan ==
> SortAggregate(key=[key#24L], functions=[max(named_struct(t1, t1#25L, key,
> key#24L, t1, t1#25L, t2, t2#26L))])
> +- SortAggregate(key=[key#24L], functions=[partial_max(named_struct(t1,
> t1#25L, key, key#24L, t1, t1#25L, t2, t2#26L))])
> +- *(1) FileScan parquet default.a1[key#24L,t1#25L,t2#26L] Batched: true,
> Format: Parquet, Location: ...{code}
>
> and here's a bad example, but more realistic:
> {code:java}
> sparkSession.sql("set spark.sql.shuffle.partitions=2")
> sparkSession.sql("select max(struct(t1, *)) from a1 group by key").explain
> == Physical Plan ==
> SortAggregate(key=[key#32L], functions=[max(named_struct(t1, t1#33L, key,
> key#32L, t1, t1#33L, t2, t2#34L))])
> +- SortAggregate(key=[key#32L], functions=[partial_max(named_struct(t1,
> t1#33L, key, key#32L, t1, t1#33L, t2, t2#34L))])
> +- *(1) Sort [key#32L ASC NULLS FIRST], false, 0
> +- *(1) FileScan parquet default.a1[key#32L,t1#33L,t2#34L] Batched: true,
> Format: Parquet, Location: ...
> {code}
>
> I've traced the problem to DataSourceScanExec#235:
> {code:java}
> val sortOrder = if (sortColumns.nonEmpty) {
> // In case of bucketing, its possible to have multiple files belonging to
> the
> // same bucket in a given relation. Each of these files are locally sorted
> // but those files combined together are not globally sorted. Given that,
> // the RDD partition will not be sorted even if the relation has sort
> columns set
> // Current solution is to check if all the buckets have a single file in it
> val files = selectedPartitions.flatMap(partition => partition.files)
> val bucketToFilesGrouping =
> files.map(_.getPath.getName).groupBy(file =>
> BucketingUtils.getBucketId(file))
> val singleFilePartitions = bucketToFilesGrouping.forall(p => p._2.length <=
> 1){code}
> so obviously the code avoids dealing with this situation now..
> could you think of a way to solve this or bypass it?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
