You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Yuming Wang (Jira)" <ji...@apache.org> on 2022/10/18 07:38:00 UTC
[jira] [Comment Edited] (SPARK-34966) Avoid shuffle if join type do not match
[ https://issues.apache.org/jira/browse/SPARK-34966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17315591#comment-17315591 ]
Yuming Wang edited comment on SPARK-34966 at 10/18/22 7:37 AM:
---------------------------------------------------------------
https://github.com/apache/spark/blob/69aa727ff495f6698fe9b37e952dfaf36f1dd5eb/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/hash.scala#L504-L507
It is same:
import org.apache.spark.sql.catalyst.expressions.{Cast, EmptyRow, Literal, Murmur3Hash, Pmod}
import org.apache.spark.sql.types.LongType
println(Pmod(new Murmur3Hash(Seq(Literal(100.toShort))), Literal(10)).eval(EmptyRow))
println(Pmod(new Murmur3Hash(Seq(new Cast(Literal(100.toShort), IntegerType))), Literal(10)).eval(EmptyRow))
But it is not same:
println(Pmod(new Murmur3Hash(Seq(Literal(100))), Literal(10)).eval(EmptyRow))
println(Pmod(new Murmur3Hash(Seq(new Cast(Literal(100), LongType))), Literal(10)).eval(EmptyRow))
was (Author: q79969786):
https://github.com/apache/spark/blob/69aa727ff495f6698fe9b37e952dfaf36f1dd5eb/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/hash.scala#L504-L507
It is same:
import org.apache.spark.sql.catalyst.expressions.{Cast, EmptyRow, Literal, Murmur3Hash, Pmod}
println(Pmod(new Murmur3Hash(Seq(Literal(100.toShort))), Literal(10)).eval(EmptyRow))
println(Pmod(new Murmur3Hash(Seq(new Cast(Literal(100.toShort), IntegerType))), Literal(10)).eval(EmptyRow))
But it is not same:
println(Pmod(new Murmur3Hash(Seq(Literal(100))), Literal(10)).eval(EmptyRow))
println(Pmod(new Murmur3Hash(Seq(new Cast(Literal(100), LongType))), Literal(10)).eval(EmptyRow))
> Avoid shuffle if join type do not match
> ---------------------------------------
>
> Key: SPARK-34966
> URL: https://issues.apache.org/jira/browse/SPARK-34966
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 3.2.0
> Reporter: Yuming Wang
> Priority: Major
>
> How to reproduce this issue:
> {code:scala}
> spark.sql("set spark.sql.autoBroadcastJoinThreshold=-1")
> spark.sql("CREATE TABLE t1 using parquet clustered by (id) into 200 buckets AS SELECT cast(id as bigint) FROM range(1000)")
> spark.sql("CREATE TABLE t2 using parquet clustered by (id) into 200 buckets AS SELECT cast(id as int) FROM range(500)")
> spark.sql("select * from t1 join t2 on (t1.id = t2.id)").explain
> {code}
> Current:
> {noformat}
> == Physical Plan ==
> AdaptiveSparkPlan isFinalPlan=false
> +- SortMergeJoin [id#14L], [cast(id#15 as bigint)], Inner
> :- Sort [id#14L ASC NULLS FIRST], false, 0
> : +- Filter isnotnull(id#14L)
> : +- FileScan parquet default.t1[id#14L] Batched: true, DataFilters: [isnotnull(id#14L)], Format: Parquet, Location: InMemoryFileIndex(1 paths)[file:/Users/yumwang/opensource/spark/spark-warehouse/org.apache.spark...., PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint>, SelectedBucketsCount: 200 out of 200
> +- Sort [cast(id#15 as bigint) ASC NULLS FIRST], false, 0
> +- Exchange hashpartitioning(cast(id#15 as bigint), 200), ENSURE_REQUIREMENTS, [id=#58]
> +- Filter isnotnull(id#15)
> +- FileScan parquet default.t2[id#15] Batched: true, DataFilters: [isnotnull(id#15)], Format: Parquet, Location: InMemoryFileIndex(1 paths)[file:/Users/yumwang/opensource/spark/spark-warehouse/org.apache.spark...., PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:int>
> {noformat}
> Expected:
> {noformat}
> == Physical Plan ==
> AdaptiveSparkPlan isFinalPlan=false
> +- SortMergeJoin [id#14L], [cast(id#15 as bigint)], Inner
> :- Sort [id#14L ASC NULLS FIRST], false, 0
> : +- Filter isnotnull(id#14L)
> : +- FileScan parquet default.t1[id#14L] Batched: true, DataFilters: [isnotnull(id#14L)], Format: Parquet, Location: InMemoryFileIndex(1 paths)[file:/Users/yumwang/opensource/spark/spark-warehouse/org.apache.spark...., PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint>, SelectedBucketsCount: 200 out of 200
> +- Sort [cast(id#15 as bigint) ASC NULLS FIRST], false, 0
> +- Filter isnotnull(id#15)
> +- FileScan parquet default.t2[id#15] Batched: true, DataFilters: [isnotnull(id#15)], Format: Parquet, Location: InMemoryFileIndex(1 paths)[file:/Users/yumwang/opensource/spark/spark-warehouse/org.apache.spark...., PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:int>, SelectedBucketsCount: 200 out of 200
> {noformat}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org