You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "psavalle (via GitHub)" <gi...@apache.org> on 2026/08/21 09:41:58 UTC
[PR] [SPARK-47672] Mark PythonUDF expensive to avoid double evaluation [spark]
psavalle opened a new pull request, #58201:
URL: https://github.com/apache/spark/pull/58201
<!--
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
'common/utils/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
'common/utils/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.
-->
Mark Python UDFs `expensive` to avoid double evaluation with filter pushdown.
### 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.
-->
https://github.com/apache/spark/pull/46143 added an `expensive` flag on expressions to avoid double evaluation with filter pushdown. The docs mention UDFs in general:
> When true avoid pushing expensive (UDF, etc.) filters down
However, only `ScalaUDF` is currently marked expensive, and the Python UDFs can still be evaluated twice.
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as new features, bug fixes, or other behavior changes. Documentation-only updates are not considered user-facing changes.
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, `spark.sql.optimizer.avoidDoubleFilterEval` now also applies to Python UDFs
### 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.
-->
### Was this patch authored or co-authored using generative AI tooling?
<!--
If generative AI tooling has been used in the process of authoring this patch, please include the
phrase: 'Generated-by: ' followed by the name of the tool and its version.
If no, write 'No'.
Please refer to the [ASF Generative Tooling Guidance](https://www.apache.org/legal/generative-tooling.html) for details.
-->
Yes, using `codex-cli 0.147.0 (gpt-5.6-sol)`
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "uros-b (via GitHub)" <gi...@apache.org>.
uros-b commented on code in PR #58201:
URL: https://github.com/apache/spark/pull/58201#discussion_r3831091669
##########
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/FilterPushdownSuite.scala:
##########
@@ -202,6 +202,68 @@ class FilterPushdownSuite extends PlanTest {
comparePlans(optimized, correctAnswer)
}
+ test("SPARK-39481: avoid evaluating scalar Python UDFs twice") {
Review Comment:
Should this be SPARK-39481 or SPARK-47672 (as per the PR title)?
##########
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/FilterPushdownSuite.scala:
##########
@@ -202,6 +202,68 @@ class FilterPushdownSuite extends PlanTest {
comparePlans(optimized, correctAnswer)
}
+ test("SPARK-39481: avoid evaluating scalar Python UDFs twice") {
+ val evalTypes = Seq(
+ "regular" -> PythonEvalType.SQL_BATCHED_UDF,
+ "Arrow-optimized" -> PythonEvalType.SQL_ARROW_BATCHED_UDF,
+ "Arrow element-wise" -> PythonEvalType.SQL_ARROW_ELEMENTWISE_UDF,
+ "scalar Pandas element-wise" -> PythonEvalType.SQL_SCALAR_PANDAS_ELEMENTWISE_UDF,
+ "scalar Pandas iterator element-wise" ->
+ PythonEvalType.SQL_SCALAR_PANDAS_ITER_ELEMENTWISE_UDF,
+ "scalar Arrow element-wise" -> PythonEvalType.SQL_SCALAR_ARROW_ELEMENTWISE_UDF,
+ "scalar Arrow iterator element-wise" ->
+ PythonEvalType.SQL_SCALAR_ARROW_ITER_ELEMENTWISE_UDF,
+ "scalar Pandas" -> PythonEvalType.SQL_SCALAR_PANDAS_UDF,
+ "scalar Pandas iterator" -> PythonEvalType.SQL_SCALAR_PANDAS_ITER_UDF,
+ "scalar Arrow" -> PythonEvalType.SQL_SCALAR_ARROW_UDF,
+ "scalar Arrow iterator" -> PythonEvalType.SQL_SCALAR_ARROW_ITER_UDF)
+
+ evalTypes.foreach { case (name, evalType) =>
+ withClue(s"$name Python UDF: ") {
+ val pythonUDF = PythonUDF(
+ "pythonUDF",
+ null,
+ BooleanType,
+ Seq(attrA),
+ evalType,
+ udfDeterministic = true)
+ val originalQuery = testRelation
+ .select(pythonUDF.as("result"))
+ .where($"result")
+ .analyze
+
+ comparePlans(Optimize.execute(originalQuery), originalQuery)
+ }
+ }
+ }
+
+ test("SPARK-39481: all Python function expressions are expensive") {
Review Comment:
Should this be SPARK-39481 or SPARK-47672 (as per the PR title)?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "Yicong-Huang (via GitHub)" <gi...@apache.org>.
Yicong-Huang commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5764073237
you would also want to update the PR title so it links to the correct JIRA ;)
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "holdenk (via GitHub)" <gi...@apache.org>.
holdenk commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5448500682
LGTM
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "psavalle (via GitHub)" <gi...@apache.org>.
psavalle commented on code in PR #58201:
URL: https://github.com/apache/spark/pull/58201#discussion_r3831274183
##########
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/FilterPushdownSuite.scala:
##########
@@ -202,6 +202,68 @@ class FilterPushdownSuite extends PlanTest {
comparePlans(optimized, correctAnswer)
}
+ test("SPARK-39481: avoid evaluating scalar Python UDFs twice") {
Review Comment:
Good point, I was looking at both, let me use SPARK-47672 everywhere since that is the one for the main mechanism (and we address not just pandas, but all Python UDFs).
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "psavalle (via GitHub)" <gi...@apache.org>.
psavalle commented on code in PR #58201:
URL: https://github.com/apache/spark/pull/58201#discussion_r3831279624
##########
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/FilterPushdownSuite.scala:
##########
@@ -202,6 +202,68 @@ class FilterPushdownSuite extends PlanTest {
comparePlans(optimized, correctAnswer)
}
+ test("SPARK-39481: avoid evaluating scalar Python UDFs twice") {
Review Comment:
Done!
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "psavalle (via GitHub)" <gi...@apache.org>.
psavalle commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5757347930
@cloud-fan can this be merged? Thanks!
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "psavalle (via GitHub)" <gi...@apache.org>.
psavalle commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5440989321
@cloud-fan thanks, done!
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-59693][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "Yicong-Huang (via GitHub)" <gi...@apache.org>.
Yicong-Huang commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5769143168
**Merge Summary:**
- merged into master https://github.com/apache/spark/commit/7fcfe8fc5709f4b0da902f5c71f8bee04f7ebbe4
- merged into branch-4.x https://github.com/apache/spark/commit/97400a43472fac800dfacaed4fff76fb56281995
*Posted by `merge_spark_pr.py`*
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5433780234
can you rebase with latest master and fix the CI jobs?
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-59693][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "Yicong-Huang (via GitHub)" <gi...@apache.org>.
Yicong-Huang closed pull request #58201: [SPARK-59693][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation
URL: https://github.com/apache/spark/pull/58201
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "psavalle (via GitHub)" <gi...@apache.org>.
psavalle commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5427584584
Oops, I had forgotten to push after updating the issue ID in the tests. Done, and updated the PR description as well. Thanks for taking a look!
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "psavalle (via GitHub)" <gi...@apache.org>.
psavalle commented on code in PR #58201:
URL: https://github.com/apache/spark/pull/58201#discussion_r3831285036
##########
sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/optimizer/FilterPushdownSuite.scala:
##########
@@ -202,6 +202,68 @@ class FilterPushdownSuite extends PlanTest {
comparePlans(optimized, correctAnswer)
}
+ test("SPARK-39481: avoid evaluating scalar Python UDFs twice") {
+ val evalTypes = Seq(
+ "regular" -> PythonEvalType.SQL_BATCHED_UDF,
+ "Arrow-optimized" -> PythonEvalType.SQL_ARROW_BATCHED_UDF,
+ "Arrow element-wise" -> PythonEvalType.SQL_ARROW_ELEMENTWISE_UDF,
+ "scalar Pandas element-wise" -> PythonEvalType.SQL_SCALAR_PANDAS_ELEMENTWISE_UDF,
+ "scalar Pandas iterator element-wise" ->
+ PythonEvalType.SQL_SCALAR_PANDAS_ITER_ELEMENTWISE_UDF,
+ "scalar Arrow element-wise" -> PythonEvalType.SQL_SCALAR_ARROW_ELEMENTWISE_UDF,
+ "scalar Arrow iterator element-wise" ->
+ PythonEvalType.SQL_SCALAR_ARROW_ITER_ELEMENTWISE_UDF,
+ "scalar Pandas" -> PythonEvalType.SQL_SCALAR_PANDAS_UDF,
+ "scalar Pandas iterator" -> PythonEvalType.SQL_SCALAR_PANDAS_ITER_UDF,
+ "scalar Arrow" -> PythonEvalType.SQL_SCALAR_ARROW_UDF,
+ "scalar Arrow iterator" -> PythonEvalType.SQL_SCALAR_ARROW_ITER_UDF)
+
+ evalTypes.foreach { case (name, evalType) =>
+ withClue(s"$name Python UDF: ") {
+ val pythonUDF = PythonUDF(
+ "pythonUDF",
+ null,
+ BooleanType,
+ Seq(attrA),
+ evalType,
+ udfDeterministic = true)
+ val originalQuery = testRelation
+ .select(pythonUDF.as("result"))
+ .where($"result")
+ .analyze
+
+ comparePlans(Optimize.execute(originalQuery), originalQuery)
+ }
+ }
+ }
+
+ test("SPARK-39481: all Python function expressions are expensive") {
Review Comment:
Updated as well.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "Yicong-Huang (via GitHub)" <gi...@apache.org>.
Yicong-Huang commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5763796770
I will merge this. thanks @psavalle
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "Yicong-Huang (via GitHub)" <gi...@apache.org>.
Yicong-Huang commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5763870612
A question though, before merging: I found the ticket SPARK-47672 was resolved to version 4.2. For this PR, it reuses the ticket so it's more of a follow up PR. However, as 4.2 is already shipped, we can't have a follow up for the ticket. @psavalle Do you want to obtain a new ticket and re-bind this PR? It will be treated as a new feature and targeting 4.x (which will be 4.4).
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47672][SQL][PYTHON] Mark PythonUDF expensive to avoid double evaluation [spark]
Posted by "psavalle (via GitHub)" <gi...@apache.org>.
psavalle commented on PR #58201:
URL: https://github.com/apache/spark/pull/58201#issuecomment-5764062492
Sure, I have filed https://issues.apache.org/jira/browse/SPARK-59693 and updated the references here
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org