rich7420 opened a new pull request, #6881: URL: https://github.com/apache/datafusion-comet/pull/6881
## Which issue does this PR close? Part of #6634. Decimal, nested hash, required-child and bitwise migration remain separate. This branch is independent of #6864, which migrates the required-child regressions. ## Rationale for this change Move scalar hash coverage into the SQL fixture framework so native execution and Spark parity can be checked without duplicate Scala setup. This slice replaces 18 Scala tests while preserving their input types, dictionary settings and child order. ## What changes are included in this PR? - Add typed Parquet fixtures for boolean, integer, binary, date and timestamp hash inputs, including boundaries and NULLs. Add TIMESTAMP_NTZ and pre-epoch coverage. - Extend existing hash/CRC32 fixtures with dictionary settings, STRING and DOUBLE cases, additional hash children, and a fixed 1,000-row boundary corpus with partially NULL inputs and control characters. Retain literal SHA2, signed-zero and NaN regressions. - Remove the 18 corresponding Scala tests. Retain hash fuzz, decimal, nested, required-child and bitwise tests. No production changes. ## How are these changes tested? Fresh native build and root-reactor Spark 4.1.3 / JDK 21 validation: - CometSqlFileTestSuite expressions/hash/: 5 tests passed. - CometHashExpressionSuite: 28 tests passed, including retained fuzz coverage. - CometExpressionSuite hash: 2 tests passed. - Compilation, Spotless, Scalastyle and the repository CI RAT script passed. A separate local probe verified 1,000 rows in both configurations, independent NULLs, preserved control characters and negative zero, and dictionary-encoded Parquet pages only when dictionary writing is enabled. This does not claim dictionary arrays survive to the hash kernels. Before queueing, run the remaining Spark profiles with `run-all-spark-profiles`. Those profiles have not yet been validated for 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]
