Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/18159#discussion_r125803128
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileFormatWriter.scala
---
@@ -314,21 +339,40 @@ object FileFormatWriter extends Logging {
recordsInFile = 0
releaseResources()
+ numOutputRows += recordsInFile
newOutputWriter(fileCounter)
}
val internalRow = iter.next()
+ val startTime = System.nanoTime()
currentWriter.write(internalRow)
+ timeOnCurrentFile += (System.nanoTime() - startTime)
--- End diff --
I'm neutral about the per-row time tracking. Comparing with actual data
writing, it should be faster. IMHO, so the performance penalty may be ignored?
In both `SingleDirectoryWriteTask` and `DynamicPartitionWriteTask`, in each
iteration, we first pull the row from the iterator and then write the row out
(and measure the timing). So the time spent on pulling rows from data pipeline
should be already excluded. That's why I'm not worried it may be inaccurate due
to data pipelining.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
