sunchao opened a new pull request, #6427: URL: https://github.com/apache/datafusion-comet/pull/6427
## Which issue does this PR close? No linked issue. ## Rationale for this change After a runtime Bloom subquery becomes a binary literal, rendering `CometFilterExec` for EXPLAIN or the Spark SQL UI can hex-encode the entire Bloom filter. A 1 MiB filter becomes roughly 2 MiB of text, even though its byte contents are not useful in a plan display. ## What changes are included in this PR? Summarize binary literals within the Bloom-filter operand as `<bloom: N bytes>` before formatting the condition. Both ordinary plan strings and formatted EXPLAIN use this display expression. The executable condition and serialized native operator retain their original values, and rendering does not evaluate unresolved subqueries. ## How are these changes tested? Four regression tests cover Spark plan strings, `SparkPlanInfo`, SQL UI graph rendering, nested predicates, ordinary binary literals, empty/null Bloom filters, and unresolved subqueries. The large-payload test throws if rendering formats or evaluates the original literal. The new suite is registered in both Linux and macOS CI matrices. Validation on Spark 4.1.3 / Scala 2.13.17 / JDK 21: - The new suite fails three of four tests on base `aca67fd581fb1838a2745f5676b49f4deee3bb8d`; all four pass with this change. - `./mvnw -Pspark-4.1 test -Dtest=none -Dsuites=org.apache.spark.sql.execution.CometFilterDisplaySuite` passes from the repository root, using the native library built from the same OSS base (native code is unchanged). - Root-reactor Spotless, `dev/ci/check-suites.py`, `dev/ci/check-ci-config.py`, and `git diff --check` pass. -- 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]
