rich7420 opened a new pull request, #6882:
URL: https://github.com/apache/datafusion-comet/pull/6882

   ## Which issue does this PR close?
   
   Part of #6627. Other regex categories remain separate migrations. This 
branch is independent of #6863, which migrates Java RLIKE tests.
   
   ## Rationale for this change
   
   Move regexp_extract, regexp_extract_all and regexp_instr coverage into SQL 
fixtures while preserving the original Java regex semantics and input cases. 
The fixtures explicitly assert JVM codegen dispatch and compare the results 
with Spark.
   
   ## What changes are included in this PR?
   
   - Extend the three existing default fixtures with the original subjects, 
patterns and group arguments, including lookahead, backreferences, embedded 
flags, no-match cases and a separate all-NULL extract table. Include subjects 
in results to preserve their association with each output.
   - Add a 5,000-row extract fixture with one Parquet write partition, batch 
size 64 and explicit row IDs.
   - Remove the 18 corresponding Scala tests. Preserve RLIKE, replace/split, 
combined-expression, error/stress and collation tests, as well as native opt-in 
fixtures. No production changes.
   
   ## How are these changes tested?
   
   Spark 4.1.3 / Scala 2.13.17 / JDK 21, root-reactor build with native JNI 
whose source and checksum match current main:
   
   - Extract SQL fixtures: 5 passed, including unchanged native opt-in fixtures.
   - Instr SQL fixture: 1 passed.
   - Remaining CometRegExpJvmSuite: 55 passed, 0 failures. One unchanged full 
RLIKE reproducer was canceled because it requires Spark 4.0 / JDK 21.
   - A separate local probe verified one input/output partition, one Parquet 
file, 79 actual Arrow batches (78 x 64 + 8), 5,000 row identities and results, 
and 715 NULL / 4,285 non-NULL extractions.
   - Compilation, Spotless, Scalastyle, diff checks and the repository CI RAT 
script passed.
   
   No migrated query is skipped. The unchanged opt-in fixtures retain their two 
pre-existing ignored lookahead queries.
   
   Run the remaining supported Spark profiles with `run-all-spark-profiles` 
before queueing; they have not yet been verified on this branch.
   


-- 
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]

Reply via email to