Skip to content

[SPARK-17972][SQL] Build Datasets upon withCachedData instead of analyzed to avoid slow query planning #15517

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
liancheng wants to merge 2 commits into from
Closed

[SPARK-17972][SQL] Build Datasets upon withCachedData instead of analyzed to avoid slow query planning #15517

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
e1283a8
Build Datasets upon cached plans
liancheng Oct 18, 2016
fffbd69
Fix streaming test failures
liancheng Oct 19, 2016
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/core/src/main/scala/org/apache/spark/sql/Dataset.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -184,7 +184,7 @@ class Dataset[T] private[sql](
case Union(children) if children.forall(hasSideEffects) =>
LogicalRDD(queryExecution.analyzed.output, queryExecution.toRdd)(sparkSession)
case _ =>
queryExecution.analyzed
queryExecution.withCachedData
}
}

Expand Down
8 changes: 6 additions & 2 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/QueryExecution.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,11 +66,13 @@ class QueryExecution(val sparkSession: SparkSession, val logical: LogicalPlan) {

lazy val withCachedData: LogicalPlan = {
assertAnalyzed()
assertSupported()
sparkSession.sharedState.cacheManager.useCachedData(analyzed)
Copy link
Contributor

@yhuai yhuai Oct 20, 2016

Choose a reason for hiding this comment

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

why?

Copy link
Contributor Author

@liancheng liancheng Oct 20, 2016

Choose a reason for hiding this comment

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

This line is actually moved to optimizedPlan. It's for fixing the streaming test failures.

Although streaming queries doesn't use QueryExecution for actual execution, it somehow triggers this line after changes made in the previous commit and throws exception. Not quite familiar with structured streaming though, I may missed something here.

}

lazy val optimizedPlan: LogicalPlan = sparkSession.sessionState.optimizer.execute(withCachedData)
lazy val optimizedPlan: LogicalPlan = {
assertSupported()
sparkSession.sessionState.optimizer.execute(withCachedData)
}

lazy val sparkPlan: SparkPlan = {
SparkSession.setActiveSession(sparkSession)
Expand Down Expand Up @@ -226,6 +228,8 @@ class QueryExecution(val sparkSession: SparkSession, val logical: LogicalPlan) {
|${stringOrError(logical.treeString(verbose = true))}
|== Analyzed Logical Plan ==
|$analyzedPlan
|== With cache data ==
|$withCachedData
|== Optimized Logical Plan ==
|${stringOrError(optimizedPlan.treeString(verbose = true))}
|== Physical Plan ==
Expand Down
2 changes: 1 addition & 1 deletion sql/core/src/main/scala/org/apache/spark/sql/execution/columnar/InMemoryRelation.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -142,7 +142,7 @@ case class InMemoryRelation(

cached.setName(
tableName.map(n => s"In-memory table $n")
.getOrElse(StringUtils.abbreviate(child.toString, 1024)))
.getOrElse(StringUtils.abbreviate(child.simpleString, 1024)))
Copy link
Contributor Author

@liancheng liancheng Oct 17, 2016

Choose a reason for hiding this comment

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

The original toString method may OOM for super large query plans. This is especially true for those plan trees built in iterative manner and grow exponentially.

_cachedColumnBuffers = cached
}

Expand Down