Skip to content

[SPARK-26450][SQL] Avoid rebuilding map of schema for every column in projection #23392

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
bersprockets wants to merge 12 commits into from
Closed

[SPARK-26450][SQL] Avoid rebuilding map of schema for every column in projection #23392

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
6b66711
Initial commit
bersprockets Dec 12, 2018
a9dbbec
Initial try at helper function
bersprockets Dec 29, 2018
67a943a
Use helper function
bersprockets Dec 31, 2018
3362744
More uses of helper function
bersprockets Dec 31, 2018
306e0a5
Even more uses of helper function
bersprockets Dec 31, 2018
a25b59c
Some clean up
bersprockets Jan 2, 2019
b977d3e
Review feedback
bersprockets Jan 2, 2019
1497d3a
Fix style issue
bersprockets Jan 2, 2019
777c5b4
Move helper function to BindReferences
bersprockets Jan 4, 2019
fbd6787
Handle cases missed before
bersprockets Jan 4, 2019
686e7b5
Review feedback
bersprockets Jan 5, 2019
b1afb3a
Close gap in ExpandExec
bersprockets Jan 13, 2019
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/expressions/Projection.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ import org.apache.spark.sql.types.{DataType, StructType}
*/
class InterpretedProjection(expressions: Seq[Expression]) extends Projection {
def this(expressions: Seq[Expression], inputSchema: Seq[Attribute]) =
this(expressions.map(BindReferences.bindReference(_, inputSchema)))
this(toBoundExprs(expressions, inputSchema))

override def initialize(partitionIndex: Int): Unit = {
expressions.foreach(_.foreach {
Expand Down
2 changes: 1 addition & 1 deletion ...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 @@ -35,7 +35,7 @@ object GenerateMutableProjection extends CodeGenerator[Seq[Expression], MutableP
in.map(ExpressionCanonicalizer.execute)

protected def bind(in: Seq[Expression], inputSchema: Seq[Attribute]): Seq[Expression] =
in.map(BindReferences.bindReference(_, inputSchema))
toBoundExprs(in, inputSchema)

def generate(
expressions: Seq[Expression],
Expand Down
4 changes: 2 additions & 2 deletions ...t/src/main/scala/org/apache/spark/sql/catalyst/expressions/codegen/GenerateOrdering.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,7 +46,7 @@ object GenerateOrdering extends CodeGenerator[Seq[SortOrder], Ordering[InternalR
in.map(ExpressionCanonicalizer.execute(_).asInstanceOf[SortOrder])

protected def bind(in: Seq[SortOrder], inputSchema: Seq[Attribute]): Seq[SortOrder] =
in.map(BindReferences.bindReference(_, inputSchema))
toBoundExprs(in, inputSchema)

/**
* Creates a code gen ordering for sorting this schema, in ascending order.
Expand Down Expand Up @@ -188,7 +188,7 @@ class LazilyGeneratedOrdering(val ordering: Seq[SortOrder])
extends Ordering[InternalRow] with KryoSerializable {

def this(ordering: Seq[SortOrder], inputSchema: Seq[Attribute]) =
this(ordering.map(BindReferences.bindReference(_, inputSchema)))
this(toBoundExprs(ordering, inputSchema))

@transient
private[this] var generatedOrdering = GenerateOrdering.generate(ordering)
Expand Down
2 changes: 1 addition & 1 deletion ...main/scala/org/apache/spark/sql/catalyst/expressions/codegen/GenerateSafeProjection.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,7 +41,7 @@ object GenerateSafeProjection extends CodeGenerator[Seq[Expression], Projection]
in.map(ExpressionCanonicalizer.execute)

protected def bind(in: Seq[Expression], inputSchema: Seq[Attribute]): Seq[Expression] =
in.map(BindReferences.bindReference(_, inputSchema))
toBoundExprs(in, inputSchema)

private def createCodeForStruct(
ctx: CodegenContext,
Expand Down
2 changes: 1 addition & 1 deletion ...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 @@ -317,7 +317,7 @@ object GenerateUnsafeProjection extends CodeGenerator[Seq[Expression], UnsafePro
in.map(ExpressionCanonicalizer.execute)

protected def bind(in: Seq[Expression], inputSchema: Seq[Attribute]): Seq[Expression] =
in.map(BindReferences.bindReference(_, inputSchema))
toBoundExprs(in, inputSchema)

def generate(
expressions: Seq[Expression],
Expand Down
2 changes: 1 addition & 1 deletion sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/ordering.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,7 @@ import org.apache.spark.sql.types._
class InterpretedOrdering(ordering: Seq[SortOrder]) extends Ordering[InternalRow] {

def this(ordering: Seq[SortOrder], inputSchema: Seq[Attribute]) =
this(ordering.map(BindReferences.bindReference(_, inputSchema)))
this(toBoundExprs(ordering, inputSchema))

def compare(a: InternalRow, b: InternalRow): Int = {
var i = 0
Expand Down
4 changes: 3 additions & 1 deletion sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/package.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,9 @@ package object expressions {
/**
* A helper function to bind given expressions to an input schema.
*/
def toBoundExprs(exprs: Seq[Expression], inputSchema: Seq[Attribute]): Seq[Expression] = {
def toBoundExprs[A <: Expression](
Copy link
Contributor

@hvanhovell hvanhovell Jan 3, 2019

Choose a reason for hiding this comment

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

I would like to minimize the chance that future changes suffer from the same issue. In order to do that we should provide API in a logical place, it does not make a whole lot of sense to me that I need to look in package.scala to find a more performant version of BindReferences.bindReference(..) for a seq. Can we move this function to BindReference and name it bindReferences?

exprs: Seq[A],
Copy link
Contributor

@mgaido91 mgaido91 Jan 2, 2019

Choose a reason for hiding this comment

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

nit: indent

inputSchema: AttributeSeq): Seq[A] = {
exprs.map(BindReferences.bindReference(_, inputSchema))
}

Expand Down
6 changes: 3 additions & 3 deletions 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 @@ -264,7 +264,7 @@ case class HashAggregateExec(
}
}
ctx.currentVars = bufVars ++ input
val boundUpdateExpr = updateExpr.map(BindReferences.bindReference(_, inputAttrs))
val boundUpdateExpr = toBoundExprs(updateExpr, inputAttrs)
val subExprs = ctx.subexpressionEliminationForWholeStageCodegen(boundUpdateExpr)
val effectiveCodes = subExprs.codes.mkString("\n")
val aggVals = ctx.withSubExprEliminationExprs(subExprs.states) {
Expand Down Expand Up @@ -825,7 +825,7 @@ case class HashAggregateExec(

val updateRowInRegularHashMap: String = {
ctx.INPUT_ROW = unsafeRowBuffer
val boundUpdateExpr = updateExpr.map(BindReferences.bindReference(_, inputAttr))
val boundUpdateExpr = toBoundExprs(updateExpr, inputAttr)
val subExprs = ctx.subexpressionEliminationForWholeStageCodegen(boundUpdateExpr)
val effectiveCodes = subExprs.codes.mkString("\n")
val unsafeRowBufferEvals = ctx.withSubExprEliminationExprs(subExprs.states) {
Expand All @@ -849,7 +849,7 @@ case class HashAggregateExec(
if (isFastHashMapEnabled) {
if (isVectorizedHashMapEnabled) {
ctx.INPUT_ROW = fastRowBuffer
val boundUpdateExpr = updateExpr.map(BindReferences.bindReference(_, inputAttr))
val boundUpdateExpr = toBoundExprs(updateExpr, inputAttr)
val subExprs = ctx.subexpressionEliminationForWholeStageCodegen(boundUpdateExpr)
val effectiveCodes = subExprs.codes.mkString("\n")
val fastRowEvals = ctx.withSubExprEliminationExprs(subExprs.states) {
Expand Down
5 changes: 2 additions & 3 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/joins/HashJoin.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,9 +63,8 @@ trait HashJoin {
protected lazy val (buildKeys, streamedKeys) = {
require(leftKeys.map(_.dataType) == rightKeys.map(_.dataType),
"Join keys from two sides should have same types")
val lkeys = HashJoin.rewriteKeyExpr(leftKeys).map(BindReferences.bindReference(_, left.output))
val rkeys = HashJoin.rewriteKeyExpr(rightKeys)
.map(BindReferences.bindReference(_, right.output))
val lkeys = toBoundExprs(HashJoin.rewriteKeyExpr(leftKeys), left.output)
val rkeys = toBoundExprs(HashJoin.rewriteKeyExpr(rightKeys), right.output)
buildSide match {
case BuildLeft => (lkeys, rkeys)
case BuildRight => (rkeys, lkeys)
Expand Down
2 changes: 1 addition & 1 deletion sql/core/src/main/scala/org/apache/spark/sql/execution/joins/SortMergeJoinExec.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -393,7 +393,7 @@ case class SortMergeJoinExec(
input: Seq[Attribute]): Seq[ExprCode] = {
ctx.INPUT_ROW = row
ctx.currentVars = null
keys.map(BindReferences.bindReference(_, input).genCode(ctx))
toBoundExprs(keys, input).map(_.genCode(ctx))
}

private def copyKeys(ctx: CodegenContext, vars: Seq[ExprCode]): Seq[ExprCode] = {
Expand Down