Skip to content

[SPARK-53225][SQL] Upgrade datasketches-java to 7.0.1 and datasketches-memory to 4.1.0#55943

Closed
yadavay-amzn wants to merge 1 commit into
apache:masterfrom
yadavay-amzn:fix/SPARK-53225-datasketches-upgrade
Closed

[SPARK-53225][SQL] Upgrade datasketches-java to 7.0.1 and datasketches-memory to 4.1.0#55943
yadavay-amzn wants to merge 1 commit into
apache:masterfrom
yadavay-amzn:fix/SPARK-53225-datasketches-upgrade

Conversation

@yadavay-amzn
Copy link
Copy Markdown
Contributor

What changes were proposed in this pull request?

Upgrades datasketches-java from 6.2.0 to 7.0.1 and adds explicit datasketches-memory 4.1.0 dependency.

Key changes:

  • pom.xml: version bumps + explicit memory dependency
  • ThetaSketchUtils.scala / TupleSketchUtils.scala: replaced removed MemoryBoundsException with IllegalArgumentException
  • Removed ResourceImpl.java (SPARK-53327 workaround, incompatible with memory 4.1.0)
  • Removed workaround references from make-distribution.sh, build_api_docs.rb, checkstyle-suppressions.xml

Why are the changes needed?

Part of Spark 4.3 release work (SPARK-56754). The new version includes bug fixes and performance improvements. Also removes the SPARK-53327 workaround which is no longer needed.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

All catalyst tests pass (8926 tests), including DatasketchesHllSketchSuite, ThetasketchesAggSuite, ThetaSketchUtilsSuite, TupleSketchUtilsSuite.

Was this patch authored or co-authored using generative AI tooling?

Yes.

…s-memory to 4.1.0

This upgrades datasketches-java from 6.2.0 to 7.0.1 and pins datasketches-memory at 4.1.0.

Changes:
- Update datasketches.version property from 6.2.0 to 7.0.1
- Add explicit datasketches-memory 4.1.0 dependency in dependencyManagement
- Replace MemoryBoundsException (removed in memory 4.x) with IllegalArgumentException
- Add jdk.incubator.foreign to test JVM modules (required by memory 4.1.0)
- Remove SPARK-53327 Java 25 workaround (incompatible with memory 4.1.0 API)

Closes #SPARK-53225
@yadavay-amzn
Copy link
Copy Markdown
Contributor Author

Closing -- @dongjoon-hyun had a prior PR (#51951) for this that was closed due to Java 17/21 distribution complexity. Deferring to his work on this.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant