Skip to content

[SPARK-37001][SQL] Disable two level of map for final hash aggregation by default #34270

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
c21 wants to merge 3 commits into from
Closed

[SPARK-37001][SQL] Disable two level of map for final hash aggregation by default #34270

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
255daac
Disable two level map for final hash aggregation by default
c21 Oct 13, 2021
d02f108
Address all comments
c21 Oct 13, 2021
889adbe
Address all comments and fix unit test failure
c21 Oct 14, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1710,6 +1710,16 @@ object SQLConf {
.booleanConf
.createWithDefault(true)

val ENABLE_TWOLEVEL_AGG_MAP_PARTIAL_ONLY =
buildConf("spark.sql.codegen.aggregate.map.twolevel.partialOnly")
.internal()
.doc("Enable two-level aggregate hash map for partial aggregate only, " +
"because final aggregate might get more distinct keys compared to partial aggregate. " +
"Overhead of looking up 1st-level map might dominate when having a lot of distinct keys.")
.version("3.2.1")
.booleanConf
.createWithDefault(true)

val ENABLE_VECTORIZED_HASH_MAP =
buildConf("spark.sql.codegen.aggregate.map.vectorized.enable")
.internal()
Expand Down Expand Up @@ -3865,6 +3875,8 @@ class SQLConf extends Serializable with Logging {

def enableTwoLevelAggMap: Boolean = getConf(ENABLE_TWOLEVEL_AGG_MAP)

def enableTwoLevelAggMapPartialOnly: Boolean = getConf(ENABLE_TWOLEVEL_AGG_MAP_PARTIAL_ONLY)
Copy link
Contributor

@cloud-fan cloud-fan Oct 14, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: we don't need to add a corresponding conf method if it's only called once.

Copy link
Contributor Author

@c21 c21 Oct 14, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@cloud-fan - updated.


def enableVectorizedHashMap: Boolean = getConf(ENABLE_VECTORIZED_HASH_MAP)

def useObjectHashAggregation: Boolean = getConf(USE_OBJECT_HASH_AGG)
Expand Down
9 changes: 8 additions & 1 deletion sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/HashAggregateExec.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -667,7 +667,14 @@ case class HashAggregateExec(
val isNotByteArrayDecimalType = bufferSchema.map(_.dataType).filter(_.isInstanceOf[DecimalType])
.forall(!DecimalType.isByteArrayDecimalType(_))

isSupported && isNotByteArrayDecimalType
val isEnabledForAggModes =
if (modes.forall(mode => mode == Partial || mode == PartialMerge)) {
true
} else {
!conf.enableTwoLevelAggMapPartialOnly
}

isSupported && isNotByteArrayDecimalType && isEnabledForAggModes
}

private def enableTwoLevelHashMap(ctx: CodegenContext): Unit = {
Expand Down