viirya commented on a change in pull request #23731: [SPARK-26572][SQL] fix
aggregate codegen result evaluation
URL: https://github.com/apache/spark/pull/23731#discussion_r255517130
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/HashAggregateExec.scala
##########
@@ -497,19 +499,25 @@ case class HashAggregateExec(
val resultVars = bindReferences[Expression](
resultExpressions,
inputAttrs).map(_.genCode(ctx))
+ val evaluateResultVars = evaluateVariables(resultVars)
s"""
$evaluateKeyVars
$evaluateResultBufferVars
+ $evaluateResultVars
${consume(ctx, resultVars)}
"""
} else {
// generate result based on grouping key
ctx.INPUT_ROW = keyTerm
ctx.currentVars = null
- val eval = bindReferences[Expression](
+ val resultVars = bindReferences[Expression](
resultExpressions,
groupingAttributes).map(_.genCode(ctx))
- consume(ctx, eval)
+ val evaluateResultVars = evaluateVariables(resultVars)
+ s"""
+ $evaluateResultVars
Review comment:
For non broadcast join cases, the change will force evaluation unnecessarily
too. We should move evaluation out of the loop in broadcast join, if possible.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
