rednaxelafx opened a new pull request, #37823:
URL: https://github.com/apache/spark/pull/37823
<!--
Thanks for sending a pull request! Here are some tips for you:
1. If this is your first time, please read our contributor guidelines:
https://spark.apache.org/contributing.html
2. Ensure you have added or run the appropriate tests for your PR:
https://spark.apache.org/developer-tools.html
3. If the PR is unfinished, add '[WIP]' in your PR title, e.g.,
'[WIP][SPARK-XXXX] Your PR title ...'.
4. Be sure to keep the PR description updated to reflect all changes.
5. Please write your PR title to summarize what this PR proposes.
6. If possible, provide a concise example to reproduce the issue for a
faster review.
7. If you want to add a new configuration, please read the guideline first
for naming configurations in
'core/src/main/scala/org/apache/spark/internal/config/ConfigEntry.scala'.
8. If you want to add or modify an error type or message, please read the
guideline first in
'core/src/main/resources/error/README.md'.
-->
### What changes were proposed in this pull request?
<!--
Please clarify what changes you are proposing. The purpose of this section
is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster
reviews in your PR. See the examples below.
1. If you refactor some codes with changing classes, showing the class
hierarchy will help reviewers.
2. If you fix some SQL features, you can provide some references of other
DBMSes.
3. If there is design documentation, please add the link.
4. If there is a discussion in the mailing list, please add the link.
-->
Block `InvokeLike` expressions with `ObjectType` result from
constant-folding, to ensure constant-folded results are trusted to be
serializable.
This is a conservative fix for ease of backport to Spark 3.3. A separate
future change can relax the restriction and support constant-folding to
serializable `ObjectType` as well.
### Why are the changes needed?
<!--
Please clarify why the changes are needed. For instance,
1. If you propose a new API, clarify the use case for a new API.
2. If you fix a bug, you can clarify why it is a bug.
-->
This fixes a regression introduced by
https://github.com/apache/spark/pull/35207 . It enabled the constant-folding
logic to aggressively fold `InvokeLike` expressions (e.g. `Invoke`,
`StaticInvoke`), when all arguments are foldable and the expression itself is
deterministic. But it could go overly aggressive and constant-fold to
non-serializable results, which would be problematic when that result needs to
be serialized and sent over the wire.
In the wild, users of sparksql-scalapb have hit this issue. The constant
folding logic would fold a chain of `Invoke` / `StaticInvoke` expressions from
only holding onto a serializable literal to holding onto a non-serializable
literal:
```
Literal(com.example.protos.demo.Person$@...).scalaDescriptor.findFieldByNumber.get
```
this expression works fine before constant-folding because the literal that
gets sent to the executors is serializable, but when aggressive
constant-folding kicks in it ends up as a
`Literal(scalapb.descriptors.FieldDescriptor@...)` which isn't serializable.
The following minimal repro demonstrates this issue:
```
import org.apache.spark.sql.Column
import org.apache.spark.sql.catalyst.analysis.UnresolvedAttribute
import org.apache.spark.sql.catalyst.expressions.Literal
import org.apache.spark.sql.catalyst.expressions.objects.{Invoke,
StaticInvoke}
import org.apache.spark.sql.types.{LongType, ObjectType}
class NotSerializableBoxedLong(longVal: Long) { def add(other: Long): Long =
longVal + other }
case class SerializableBoxedLong(longVal: Long) { def toNotSerializable():
NotSerializableBoxedLong = new NotSerializableBoxedLong(longVal) }
val litExpr = Literal.fromObject(SerializableBoxedLong(42L),
ObjectType(classOf[SerializableBoxedLong]))
val toNotSerializableExpr = Invoke(litExpr, "toNotSerializable",
ObjectType(classOf[NotSerializableBoxedLong]))
val addExpr = Invoke(toNotSerializableExpr, "add", LongType,
Seq(UnresolvedAttribute.quotedString("id")))
val df = spark.range(2).select(new Column(addExpr))
df.collect
```
would result in an error if aggressive constant-folding kicked in:
```
...
Caused by: java.io.NotSerializableException: NotSerializableBoxedLong
Serialization stack:
- object not serializable (class: NotSerializableBoxedLong, value:
NotSerializableBoxedLong@71231636)
- element of array (index: 1)
- array (class [Ljava.lang.Object;, size 2)
- element of array (index: 1)
- array (class [Ljava.lang.Object;, size 3)
- field (class: java.lang.invoke.SerializedLambda, name: capturedArgs,
type: class [Ljava.lang.Object;)
- object (class java.lang.invoke.SerializedLambda,
SerializedLambda[capturingClass=class
org.apache.spark.sql.execution.WholeStageCodegenExec,
functionalInterfaceMethod=scala/Function2.apply:(Ljava/lang/Object;Ljava/lang/Object;)Ljava/lang/Object;,
implementation=invokeStatic
org/apache/spark/sql/execution/WholeStageCodegenExec.$anonfun$doExecute$4$adapted:(Lorg/apache/spark/sql/catalyst/expressions/codegen/CodeAndComment;[Ljava/lang/Object;Lorg/apache/spark/sql/execution/metric/SQLMetric;Ljava/lang/Object;Lscala/collection/Iterator;)Lscala/collection/Iterator;,
instantiatedMethodType=(Ljava/lang/Object;Lscala/collection/Iterator;)Lscala/collection/Iterator;,
numCaptured=3])
- writeReplace data (class: java.lang.invoke.SerializedLambda)
- object (class
org.apache.spark.sql.execution.WholeStageCodegenExec$$Lambda$3123/1641694389,
org.apache.spark.sql.execution.WholeStageCodegenExec$$Lambda$3123/1641694389@185db22c)
at
org.apache.spark.serializer.SerializationDebugger$.improveException(SerializationDebugger.scala:41)
at
org.apache.spark.serializer.JavaSerializationStream.writeObject(JavaSerializer.scala:49)
at
org.apache.spark.serializer.JavaSerializerInstance.serialize(JavaSerializer.scala:115)
at
org.apache.spark.util.ClosureCleaner$.ensureSerializable(ClosureCleaner.scala:441)
```
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as
the documentation fix.
If yes, please clarify the previous behavior and the change this PR proposes
- provide the console output, description and/or an example to show the
behavior difference if possible.
If possible, please also clarify if this is a user-facing change compared to
the released Spark versions or within the unreleased branches such as master.
If no, write 'No'.
-->
Yes, a regression in ObjectType expression starting from Spark 3.3.0 is
fixed.
### How was this patch tested?
<!--
If tests were added, say they were added here. Please make sure to add some
test cases that check the changes thoroughly including negative and positive
cases if possible.
If it was tested in a way different from regular unit tests, please clarify
how you tested step by step, ideally copy and paste-able, so that other
reviewers can test and check, and descendants can verify in the future.
If tests were not added, please describe why they were not added and/or why
it was difficult to add.
If benchmark tests were added, please run the benchmarks in GitHub Actions
for the consistent environment, and the instructions could accord to:
https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
-->
The existing test cases in `ConstantFoldingSuite` continues to pass; added a
new test case to demonstrate the regression issue.
--
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]
