HyukjinKwon commented on a change in pull request #32015:
URL: https://github.com/apache/spark/pull/32015#discussion_r605461126
##########
File path:
sql/core/src/test/scala/org/apache/spark/sql/execution/SubExprEliminationBenchmark.scala
##########
@@ -45,7 +45,7 @@ object SubExprEliminationBenchmark extends SqlBasedBenchmark {
withTempPath { path =>
prepareDataInfo(benchmark)
- val numCols = 1000
+ val numCols = 500
Review comment:
@viirya, I had to reduce the number of predicates .. otherwise it fails
(only) when you run it with `spark-submit` in JDK 11. It can be reproducible
without using `Benchmarks` main.
Indeterministic exceptions. Either:
```
at
org.apache.spark.sql.catalyst.trees.TreeNode.$anonfun$collectFirst$2(TreeNode.scala:231)
at
scala.collection.LinearSeqOptimized.foldLeft(LinearSeqOptimized.scala:126)
at
scala.collection.LinearSeqOptimized.foldLeft$(LinearSeqOptimized.scala:122)
at scala.collection.immutable.List.foldLeft(List.scala:89)
at
org.apache.spark.sql.catalyst.trees.TreeNode.$anonfun$collectFirst$1(TreeNode.scala:231)
at scala.Option.orElse(Option.scala:447)
at
org.apache.spark.sql.catalyst.trees.TreeNode.collectFirst(TreeNode.scala:231)
at
org.apache.spark.sql.catalyst.trees.TreeNode.$anonfun$collectFirst$3(TreeNode.scala:231)
at scala.Option.orElse(Option.scala:447)
at
org.apache.spark.sql.catalyst.trees.TreeNode.$anonfun$collectFirst$2(TreeNode.scala:231)
at
scala.collection.LinearSeqOptimized.foldLeft(LinearSeqOptimized.scala:126)
at
scala.collection.LinearSeqOptimized.foldLeft$(LinearSeqOptimized.scala:122)
```
Or
```
at
java.base/java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1679)
at
java.base/java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2464)
at
java.base/java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:2358)
at
java.base/java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:2196)
at
java.base/java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1679)
at
java.base/java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2464)
at
java.base/java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:2358)
at
java.base/java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:2196)
at
java.base/java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1679)
at
java.base/java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2464)
at
java.base/java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:2358)
```
Even with reducing, I think it shows the performance impact with the
expression elimination.
I will update all the benchmark results according to the GA builds in a
separate PR so I hope the current change is fine to you.
--
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]
