Skip to content

[SPARK-22141][SQL] Propagate empty relation before checking Cartesian products #19362

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
gengliangwang wants to merge 2 commits into from
Closed

[SPARK-22141][SQL] Propagate empty relation before checking Cartesian products #19362

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0f25a07
propagate empty relation before checking Cartesian products
gengliangwang Sep 27, 2017
b721017
test the various join types in a single test case.
gengliangwang Sep 27, 2017
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
4 changes: 2 additions & 2 deletions sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,8 +124,6 @@ abstract class Optimizer(sessionCatalog: SessionCatalog)
SimplifyCreateMapOps,
CombineConcats) ++
extendedOperatorOptimizationRules: _*) ::
Batch("Check Cartesian Products", Once,
CheckCartesianProducts) ::
Batch("Join Reorder", Once,
CostBasedJoinReorder) ::
Batch("Decimal Optimizations", fixedPoint,
Expand All @@ -136,6 +134,8 @@ abstract class Optimizer(sessionCatalog: SessionCatalog)
Batch("LocalRelation", fixedPoint,
ConvertToLocalRelation,
PropagateEmptyRelation) ::
Batch("Check Cartesian Products", Once,
Copy link
Contributor

@caneGuy caneGuy Sep 27, 2017

Choose a reason for hiding this comment

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

I think the comment of CheckCartesianProducts should also be updated and add this constrain.

Copy link
Contributor

@rxin rxin Sep 27, 2017

Choose a reason for hiding this comment

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

we should also add a comment here about the positioning ...

Copy link
Member Author

@gengliangwang gengliangwang Sep 27, 2017

Choose a reason for hiding this comment

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

Make sense. The PR is merged, what should I do now..

Copy link
Member

@gatorsmile gatorsmile Sep 27, 2017

Choose a reason for hiding this comment

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

Create a follow-up PR

Copy link
Member Author

@gengliangwang gengliangwang Sep 27, 2017

Choose a reason for hiding this comment

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

@gatorsmile I see, thanks!

CheckCartesianProducts) ::
Batch("OptimizeCodegen", Once,
OptimizeCodegen) ::
Batch("RewriteSubquery", Once,
Expand Down
24 changes: 24 additions & 0 deletions sql/core/src/test/scala/org/apache/spark/sql/JoinSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -200,6 +200,30 @@ class JoinSuite extends QueryTest with SharedSQLContext {
Nil)
}

test("inner join, propagate empty relation before checking Cartesian products") {
Copy link
Contributor

@hvanhovell hvanhovell Sep 27, 2017 •
edited
Loading

Choose a reason for hiding this comment

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

It make make the test more concise if you just test the various join types in a single test.

val x = testData2.as("x")
val y = testData2.where($"a" === 2 && !($"a" === 2)).as("y")
checkAnswer(
x.join(y).where($"x.a" === $"y.a"),
Nil)
}

test("left outer join, propagate empty relation before checking Cartesian products") {
val x = testData2.where($"a" === 2 && !($"a" === 2)).as("x")
val y = testData2.as("y")
checkAnswer(
x.join(y, Seq.empty, "left_outer"),
Nil)
}

test("right outer join, propagate empty relation before checking Cartesian products") {
val x = testData2.as("x")
val y = testData2.where($"a" === 2 && !($"a" === 2)).as("y")
checkAnswer(
x.join(y, Seq.empty, "right_outer"),
Nil)
}

test("big inner join, 4 matches per row") {
val bigData = testData.union(testData).union(testData).union(testData)
val bigDataX = bigData.as("x")
Expand Down