viirya commented on code in PR #31451:
URL: https://github.com/apache/spark/pull/31451#discussion_r936037974
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/ui/SQLAppStatusListener.scala:
##########
@@ -199,7 +202,37 @@ class SQLAppStatusListener(
}
private def aggregateMetrics(exec: LiveExecutionData): Map[Long, String] = {
- val metricTypes = exec.metrics.map { m => (m.accumulatorId, m.metricType)
}.toMap
+ val accumIds = exec.metrics.map(_.accumulatorId).toSet
+
+ val metricAggregationMap = new mutable.HashMap[String, (Array[Long],
Array[Long]) => String]()
+ val metricAggregationMethods = exec.metrics.map { m =>
+ val optClassName = CustomMetrics.parseV2CustomMetricType(m.metricType)
+ val metricAggMethod = optClassName.map { className =>
+ if (metricAggregationMap.contains(className)) {
+ metricAggregationMap(className)
+ } else {
+ // Try to initiate custom metric object
+ try {
+ val metric = Utils.loadExtensions(classOf[CustomMetric],
Seq(className), conf).head
Review Comment:
For current usage, I don't see there are other necessary args for
`CustomMetric` constructor. Besides, as an interface, we cannot define
constructor for `CustomMetric`. I think `CustomMetric` should be pretty simple
class and I'd like to keep it as simple as possible. If there are more usages
that require additional args, we can add an initialization API used to pass
some parameters. But I doubt that if it is necessary.
--
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]
