HeartSaVioR commented on a change in pull request #31369:
URL: https://github.com/apache/spark/pull/31369#discussion_r568291356
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala
##########
@@ -165,10 +165,18 @@ case class StateStoreMetrics(
object StateStoreMetrics {
def combine(allMetrics: Seq[StateStoreMetrics]): StateStoreMetrics = {
+ val distinctCustomMetrics =
allMetrics.flatMap(_.customMetrics.keys).distinct
+ val customMetrics = allMetrics.flatMap(_.customMetrics)
+ val combinedCustomMetrics = distinctCustomMetrics.map { customMetric =>
Review comment:
My understanding is that combine is only called with multiple elements,
where we expect these elements will have same custom metrics (as these are from
state store provider). Please correct me if I'm missing here.
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/state/StateStore.scala
##########
@@ -165,10 +165,18 @@ case class StateStoreMetrics(
object StateStoreMetrics {
def combine(allMetrics: Seq[StateStoreMetrics]): StateStoreMetrics = {
+ val distinctCustomMetrics =
allMetrics.flatMap(_.customMetrics.keys).distinct
+ val customMetrics = allMetrics.flatMap(_.customMetrics)
+ val combinedCustomMetrics = distinctCustomMetrics.map { customMetric =>
Review comment:
So if you assume the chance allMetrics is only having one element, then
all these calculations are unnecessary. But this method already assumes the
case we want multiple metrics to be combined, and I think we don't blindly call
this method with one element. No benefit then.
----------------------------------------------------------------
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]
