Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/8931#discussion_r42065649
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/metric/SQLMetrics.scala
---
@@ -102,20 +96,53 @@ private object LongSQLMetricParam extends
SQLMetricParam[LongSQLMetricValue, Lon
override def zero(initialValue: LongSQLMetricValue): LongSQLMetricValue
= zero
- override def zero: LongSQLMetricValue = new LongSQLMetricValue(0L)
+ override def zero: LongSQLMetricValue = new
LongSQLMetricValue(initialValue)
}
private[sql] object SQLMetrics {
- def createLongMetric(sc: SparkContext, name: String): LongSQLMetric = {
- val acc = new LongSQLMetric(name)
+ def createLongMetric(
+ sc: SparkContext,
+ name: String,
+ stringValue: Seq[Long] => String,
+ initialValue: Long): LongSQLMetric = {
+ val param = new LongSQLMetricParam(stringValue, initialValue)
+ val acc = new LongSQLMetric(name, param)
sc.cleaner.foreach(_.registerAccumulatorForCleanup(acc))
acc
}
+ def createLongMetric(sc: SparkContext, name: String): LongSQLMetric = {
+ createLongMetric(sc, name, _.sum.toString, 0L)
+ }
+
+ /**
+ * Create a metric to report the size information(including total, min,
med, max) like data size,
+ * spilled size, etc.
+ */
+ def createSizeMetric(sc: SparkContext, name: String): LongSQLMetric = {
+ val stringValue = (values: Seq[Long]) => {
+ // This is a work around for
https://issues.apache.org/jira/browse/SPARK-11013
+ // We use -1 as initial value of the accumulator, if the accumulator
is valid, we will update
+ // it at the end of task and the value will be at least 0.
+ val validValues = values.filter(_ >= 0)
+ val Seq(sum, min, med, max) = {
+ val metric = if (validValues.length == 0) {
+ Seq.fill(4)(0L)
+ } else {
+ val sorted = validValues.sorted
+ Seq(sorted.sum, sorted(0), sorted(validValues.length / 2),
sorted(validValues.length - 1))
+ }
+ metric.map(Utils.bytesToString)
+ }
+ s"\n$sum ($min, $med, $max)"
+ }
+ createLongMetric(sc, s"$name total (min, med, max)", stringValue, -1L)
--- End diff --
can you add an example of what it looks like?
```
// e.g. data size total (min, med, max): 100g (100m, 1g, 10g)
```
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
