LucaCanali commented on a change in pull request #31367:
URL: https://github.com/apache/spark/pull/31367#discussion_r566732268
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/python/BatchEvalPythonExec.scala
##########
@@ -83,12 +97,19 @@ case class BatchEvalPythonExec(udfs: Seq[PythonUDF],
resultAttrs: Seq[Attribute]
val unpickledBatch = unpickle.loads(pickedResult)
unpickledBatch.asInstanceOf[java.util.ArrayList[Any]].asScala
}.map { result =>
+ pythonNumRowsReceived += 1
+ val startTime = System.nanoTime()
if (udfs.length == 1) {
// fast path for single UDF
mutableRow(0) = fromJava(result)
+ val deltaTime = System.nanoTime() - startTime
+ pythonExecTime += deltaTime
mutableRow
} else {
- fromJava(result).asInstanceOf[InternalRow]
+ val res = fromJava(result).asInstanceOf[InternalRow]
+ val deltaTime = System.nanoTime() - startTime
+ pythonExecTime += deltaTime
Review comment:
Correct, this is just a component, actually small and could be omitted.
The most important part of the execution for BatchEvalPythonExec time is
measured in PythonUDFRunner.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
