Skip to content

[SPARK-9462][SQL] Initialize nondeterministic expressions in code gen fallback mode. #7767

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
rxin wants to merge 3 commits into from
Closed

[SPARK-9462][SQL] Initialize nondeterministic expressions in code gen fallback mode. #7767

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
bb5c334
[SPARK-9462][SQL] Initialize nondeterministic expressions in code gen…
rxin Jul 30, 2015
713ac3a
More unit tests.
rxin Jul 30, 2015
ef3e2d9
Removed println
rxin Jul 30, 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
5 changes: 5 additions & 0 deletions ...n/scala/org/apache/spark/sql/catalyst/expressions/codegen/GenerateMutableProjection.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,6 +122,11 @@ object GenerateMutableProjection extends CodeGenerator[Seq[Expression], () => Mu

logDebug(s"code for ${expressions.mkString(",")}:\n${CodeFormatter.format(code)}")

ctx.references.foreach {
case n: Nondeterministic => n.setInitialValues()
case _ =>
}

val c = compile(code)
() => {
c.generate(ctx.references.toArray).asInstanceOf[MutableProjection]
Expand Down
5 changes: 5 additions & 0 deletions .../src/main/scala/org/apache/spark/sql/catalyst/expressions/codegen/GeneratePredicate.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,6 +62,11 @@ object GeneratePredicate extends CodeGenerator[Expression, (InternalRow) => Bool

logDebug(s"Generated predicate '$predicate':\n${CodeFormatter.format(code)}")

ctx.references.foreach {
case n: Nondeterministic => n.setInitialValues()
case _ =>
}

@cloud-fan cloud-fan Jul 30, 2015

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

it should be

ctx.references.foreach(_.foreach {
  case n: Nondeterministic => n.setInitialValues()
  case _ =>
})

@cloud-fan cloud-fan Jul 30, 2015

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

and why not do initiliaztion in CodegenFallback?

@rxin rxin Jul 30, 2015

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

ah yes

val p = compile(code).generate(ctx.references.toArray).asInstanceOf[Predicate]
(r: InternalRow) => p.eval(r)
}
Expand Down
5 changes: 5 additions & 0 deletions ...src/main/scala/org/apache/spark/sql/catalyst/expressions/codegen/GenerateProjection.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -233,6 +233,11 @@ object GenerateProjection extends CodeGenerator[Seq[Expression], Projection] {
logDebug(s"MutableRow, initExprs: ${expressions.mkString(",")} code:\n" +
CodeFormatter.format(code))

ctx.references.foreach {
case n: Nondeterministic => n.setInitialValues()
case _ =>
}

compile(code).generate(ctx.references.toArray).asInstanceOf[Projection]
}
}
5 changes: 5 additions & 0 deletions ...in/scala/org/apache/spark/sql/catalyst/expressions/codegen/GenerateUnsafeProjection.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -285,6 +285,11 @@ object GenerateUnsafeProjection extends CodeGenerator[Seq[Expression], UnsafePro

logDebug(s"code for ${expressions.mkString(",")}:\n${CodeFormatter.format(code)}")

ctx.references.foreach {
case n: Nondeterministic => n.setInitialValues()
case _ =>
}

val c = compile(code)
c.generate(ctx.references.toArray).asInstanceOf[UnsafeProjection]
}
Expand Down
45 changes: 40 additions & 5 deletions ...ala/org/apache/spark/sql/catalyst/expressions/codegen/CodegenExpressionCachingSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,15 +19,39 @@ package org.apache.spark.sql.catalyst.expressions.codegen

import org.apache.spark.SparkFunSuite
import org.apache.spark.sql.catalyst.InternalRow
import org.apache.spark.sql.catalyst.expressions.{UnsafeProjection, LeafExpression}
import org.apache.spark.sql.catalyst.expressions.{Nondeterministic, UnsafeProjection, LeafExpression}
import org.apache.spark.sql.types.{BooleanType, DataType}

/**
* A test suite that makes sure code generation handles expression internally states correctly.
*/
class CodegenExpressionCachingSuite extends SparkFunSuite {

test("GenerateUnsafeProjection") {
test("GenerateUnsafeProjection should initialize expressions") {
val expr = NondeterministicExpression()
val instance = UnsafeProjection.create(Seq(expr))
assert(instance.apply(null).getBoolean(0) === false)
}

test("GenerateProjection should initialize expressions") {
val expr = NondeterministicExpression()
val instance = GenerateProjection.generate(Seq(expr))
assert(instance.apply(null).getBoolean(0) === false)
}

test("GenerateMutableProjection should initialize expressions") {
val expr = NondeterministicExpression()
val instance = GenerateMutableProjection.generate(Seq(expr))()
assert(instance.apply(null).getBoolean(0) === false)
}

test("GeneratePredicate should initialize expressions") {
val expr = NondeterministicExpression()
val instance = GeneratePredicate.generate(expr)
assert(instance.apply(null) === false)
}

test("GenerateUnsafeProjection should not share expression instances") {
val expr1 = MutableExpression()
val instance1 = UnsafeProjection.create(Seq(expr1))
assert(instance1.apply(null).getBoolean(0) === false)
Expand All @@ -39,7 +63,7 @@ class CodegenExpressionCachingSuite extends SparkFunSuite {
assert(instance2.apply(null).getBoolean(0) === true)
}

test("GenerateProjection") {
test("GenerateProjection should not share expression instances") {
val expr1 = MutableExpression()
val instance1 = GenerateProjection.generate(Seq(expr1))
assert(instance1.apply(null).getBoolean(0) === false)
Expand All @@ -51,7 +75,7 @@ class CodegenExpressionCachingSuite extends SparkFunSuite {
assert(instance2.apply(null).getBoolean(0) === true)
}

test("GenerateMutableProjection") {
test("GenerateMutableProjection should not share expression instances") {
val expr1 = MutableExpression()
val instance1 = GenerateMutableProjection.generate(Seq(expr1))()
assert(instance1.apply(null).getBoolean(0) === false)
Expand All @@ -63,7 +87,7 @@ class CodegenExpressionCachingSuite extends SparkFunSuite {
assert(instance2.apply(null).getBoolean(0) === true)
}

test("GeneratePredicate") {
test("GeneratePredicate should not share expression instances") {
val expr1 = MutableExpression()
val instance1 = GeneratePredicate.generate(expr1)
assert(instance1.apply(null) === false)
Expand All @@ -77,6 +101,17 @@ class CodegenExpressionCachingSuite extends SparkFunSuite {

}

/**
* An expression that's non-deterministic and doesn't support codegen.
*/
case class NondeterministicExpression()
extends LeafExpression with Nondeterministic with CodegenFallback {
override protected def initInternal(): Unit = {}
override protected def evalInternal(input: InternalRow): Any = false
override def nullable: Boolean = false
override def dataType: DataType = BooleanType
}


/**
* An expression with mutable state so we can change it freely in our test suite.
Expand Down