Skip to content

[SPARK-9306][SQL] Don't use SortMergeJoin when joining on unsortable columns#7645

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
viirya wants to merge 2 commits into from
Closed

[SPARK-9306][SQL] Don't use SortMergeJoin when joining on unsortable columns #7645

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
55221fa
Shouldn't use SortMergeJoin when joining on unsortable columns.
viirya Jul 24, 2015
a240707
Use forall instead of exists for readability.
viirya Jul 24, 2015
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
2 changes: 1 addition & 1 deletion sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/planning/patterns.scala
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -184,7 +184,7 @@ object PartialAggregation {
* A pattern that finds joins with equality conditions that can be evaluated using equi-join.
*/
object ExtractEquiJoinKeys extends Logging with PredicateHelper {
/** (joinType, rightKeys, leftKeys, condition, leftChild, rightChild) */
/** (joinType, leftKeys, rightKeys, condition, leftChild, rightChild) */

@viiryaviiryaJul 24, 2015

Copy link
Copy Markdown
MemberAuthor

Choose a reason for hiding this comment

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

Incorrect comment.

type ReturnType =
(JoinType, Seq[Expression], Seq[Expression], Option[Expression], LogicalPlan, LogicalPlan)

Expand Down
19 changes: 15 additions & 4 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -36,9 +36,8 @@ private[sql] abstract class SparkStrategies extends QueryPlanner[SparkPlan] {

object LeftSemiJoin extends Strategy with PredicateHelper {
def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ExtractEquiJoinKeys(LeftSemi, leftKeys, rightKeys, condition, left, right)
if sqlContext.conf.autoBroadcastJoinThreshold > 0 &&
right.statistics.sizeInBytes <= sqlContext.conf.autoBroadcastJoinThreshold =>
case ExtractEquiJoinKeys(
LeftSemi, leftKeys, rightKeys, condition, left, CanBroadcast(right)) =>

@viiryaviiryaJul 24, 2015

Copy link
Copy Markdown
MemberAuthor

Choose a reason for hiding this comment

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

We can use CanBroadcast here.

joins.BroadcastLeftSemiJoinHash(
leftKeys, rightKeys, planLater(left), planLater(right), condition) :: Nil
// Find left semi joins where at least some predicates can be evaluated by matching join keys
Expand DownExpand Up@@ -91,6 +90,18 @@ private[sql] abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
condition.map(Filter(_, broadcastHashJoin)).getOrElse(broadcastHashJoin) :: Nil
}

private[this] def isValidSort(
leftKeys: Seq[Expression],
rightKeys: Seq[Expression]): Boolean = {
leftKeys.zip(rightKeys).forall { keys =>
(keys._1.dataType, keys._2.dataType) match {
case (l: AtomicType, r: AtomicType) => true
case (NullType, NullType) => true
case _ => false
}
}
}

def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ExtractEquiJoinKeys(Inner, leftKeys, rightKeys, condition, left, CanBroadcast(right)) =>
makeBroadcastHashJoin(leftKeys, rightKeys, left, right, condition, joins.BuildRight)
Expand All@@ -101,7 +112,7 @@ private[sql] abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
// If the sort merge join option is set, we want to use sort merge join prior to hashjoin
// for now let's support inner join first, then add outer join
case ExtractEquiJoinKeys(Inner, leftKeys, rightKeys, condition, left, right)
if sqlContext.conf.sortMergeJoinEnabled =>
if sqlContext.conf.sortMergeJoinEnabled && isValidSort(leftKeys, rightKeys) =>
val mergeJoin =
joins.SortMergeJoin(leftKeys, rightKeys, planLater(left), planLater(right))
condition.map(Filter(_, mergeJoin)).getOrElse(mergeJoin) :: Nil
Expand Down
12 changes: 12 additions & 0 deletions sql/core/src/test/scala/org/apache/spark/sql/JoinSuite.scala
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -108,6 +108,18 @@ class JoinSuite extends QueryTest with BeforeAndAfterEach {
}
}

test("SortMergeJoin shouldn't work on unsortable columns") {
val SORTMERGEJOIN_ENABLED: Boolean = ctx.conf.sortMergeJoinEnabled
try {
ctx.conf.setConf(SQLConf.SORTMERGE_JOIN, true)
Seq(
("SELECT * FROM arrayData JOIN complexData ON data = a", classOf[ShuffledHashJoin])

@viiryaviiryaJul 24, 2015

Copy link
Copy Markdown
MemberAuthor

Choose a reason for hiding this comment

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

arrayData.data and complexData.a are both Seq[Int].

).foreach { case (query, joinClass) => assertJoin(query, joinClass) }
} finally {
ctx.conf.setConf(SQLConf.SORTMERGE_JOIN, SORTMERGEJOIN_ENABLED)
}
}

test("broadcasted hash join operator selection") {
ctx.cacheManager.clearCache()
ctx.sql("CACHE TABLE testData")
Expand Down