Github user rxin commented on a diff in the pull request:
https://github.com/apache/spark/pull/23207#discussion_r238836448
--- Diff: core/src/main/scala/org/apache/spark/shuffle/metrics.scala ---
@@ -50,3 +50,57 @@ private[spark] trait ShuffleWriteMetricsReporter {
private[spark] def decBytesWritten(v: Long): Unit
private[spark] def decRecordsWritten(v: Long): Unit
}
+
+
+/**
+ * A proxy class of ShuffleWriteMetricsReporter which proxy all metrics
updating to the input
+ * reporters.
+ */
+private[spark] class GroupedShuffleWriteMetricsReporter(
--- End diff --
I'd not create a general API here. Just put one in SQL similar to the read
side that also calls the default one.
It can be expensive to go through a seq for each record and bytes.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
