xuanyuanking commented on pull request #31369: URL: https://github.com/apache/spark/pull/31369#issuecomment-774634462
Sorry for pinging you again on this issue. I touched this code today and have a question: @viirya Could you provide which metric missed half of the result in your scenario? It seems that the overridden shouldn't happen in the original codebase for the custom metrics framework. Do you use some special way for adding the custom metrics? Here's some analysis: The left/rightSideMetrics passed in `combine` function are objects with adding new descriptions here: https://github.com/apache/spark/blob/e3058ba17cb4512537953eb4ded884e24ee93ba2/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/SymmetricHashJoinStateManager.scala#L316-L321 So the result of `combine` should contain two metrics object with the same name. They will finally get summed in https://github.com/apache/spark/blob/e3058ba17cb4512537953eb4ded884e24ee93ba2/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/StreamingSymmetricHashJoinExec.scala#L403-L405 ---------------------------------------------------------------- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
