snmvaughan commented on code in PR #45123:
URL: https://github.com/apache/spark/pull/45123#discussion_r1538042782
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/InsertIntoHadoopFsRelationCommand.scala:
##########
@@ -174,6 +174,7 @@ case class InsertIntoHadoopFsRelationCommand(
qualifiedOutputPath
}
+ val statsTrackers = Seq(basicWriteJobStatsTracker(hadoopConf))
Review Comment:
The intention was to capture partition statistics during the writing and
committing phases, and then report the results. As an example, Iceberg does
collect this information already, so the only necessary change is making this
visible by asking for the aggregated results which are then posted to the bus.
The existing statistics that are gathered for V1 use `SQLMetric`, which is
an `AccumulatorV2` constrained to a simple `Long` value. This is why the
change included an additional parameter which would allow for capturing the
dynamic map of partitions to their stats. I'll look to avoid this
implementation detail in any interfaces.
The V2 updates make all of this optional, by adding a new interface for
asking an implementing source to provide the aggregated results.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
