cloud-fan commented on a change in pull request #32084:
URL: https://github.com/apache/spark/pull/32084#discussion_r718092179
##########
File path:
sql/core/src/test/scala/org/apache/spark/sql/execution/adaptive/AdaptiveQueryExecSuite.scala
##########
@@ -1705,6 +1705,89 @@ class AdaptiveQueryExecSuite
}
}
+ test("SPARK-34980: Support coalesce partition through union") {
+ def checkResultPartition(
+ df: Dataset[Row],
+ numUnion: Int,
+ numShuffleReader: Int,
+ numPartition: Int): Unit = {
+ df.collect()
+ assert(collect(df.queryExecution.executedPlan) {
+ case u: UnionExec => u
+ }.size == numUnion)
+ assert(collect(df.queryExecution.executedPlan) {
+ case r: AQEShuffleReadExec => r
+ }.size === numShuffleReader)
+ assert(df.rdd.partitions.length === numPartition)
+ }
+
+ Seq(true, false).foreach { combineUnionEnabled =>
+ val combineUnionConfig = if (combineUnionEnabled) {
+ "" -> ""
Review comment:
this will set a config whose key is an empty string. I think it's safer
to do `SQLConf.OPTIMIZER_EXCLUDED_RULES.key -> ""`
##########
File path:
sql/core/src/test/scala/org/apache/spark/sql/execution/adaptive/AdaptiveQueryExecSuite.scala
##########
@@ -1705,6 +1705,89 @@ class AdaptiveQueryExecSuite
}
}
+ test("SPARK-34980: Support coalesce partition through union") {
+ def checkResultPartition(
+ df: Dataset[Row],
+ numUnion: Int,
+ numShuffleReader: Int,
+ numPartition: Int): Unit = {
+ df.collect()
+ assert(collect(df.queryExecution.executedPlan) {
+ case u: UnionExec => u
+ }.size == numUnion)
+ assert(collect(df.queryExecution.executedPlan) {
+ case r: AQEShuffleReadExec => r
+ }.size === numShuffleReader)
+ assert(df.rdd.partitions.length === numPartition)
+ }
+
+ Seq(true, false).foreach { combineUnionEnabled =>
+ val combineUnionConfig = if (combineUnionEnabled) {
+ "" -> ""
+ } else {
+ SQLConf.OPTIMIZER_EXCLUDED_RULES.key ->
+ "org.apache.spark.sql.catalyst.optimizer.CombineUnions"
+ }
+ // advisory partition size 1048576 has no special meaning, just a big
enough value
+ withSQLConf(SQLConf.ADAPTIVE_EXECUTION_ENABLED.key -> "true",
+ SQLConf.COALESCE_PARTITIONS_ENABLED.key -> "true",
+ SQLConf.ADVISORY_PARTITION_SIZE_IN_BYTES.key -> "1048576",
+ SQLConf.COALESCE_PARTITIONS_MIN_PARTITION_NUM.key -> "1",
+ SQLConf.SHUFFLE_PARTITIONS.key -> "10",
+ combineUnionConfig) {
+ withTempView("t1", "t2") {
+ spark.sparkContext.parallelize((1 to 10).map(i => TestData(i,
i.toString)), 2)
+ .toDF().createOrReplaceTempView("t1")
+ spark.sparkContext.parallelize((1 to 10).map(i => TestData(i,
i.toString)), 4)
+ .toDF().createOrReplaceTempView("t2")
+
+ // positive test that could be coalesced
+ checkResultPartition(
+ sql("""
+ |SELECT key, count(*) FROM t1 GROUP BY key
+ |UNION ALL
+ |SELECT * FROM t2
+ """.stripMargin),
+ 1,
Review comment:
can we put the parameter name to make the test more readable? `numUnion
= 1`
--
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]
