peterxcli opened a new pull request, #5177: URL: https://github.com/apache/datafusion-comet/pull/5177
## Which issue does this PR close? Closes #5090. ## Rationale for this change Several native temporal conversions duplicate Arrow kernels with hand-written per-row loops. Reusing Arrow centralizes the unit conversion and null handling while reducing code to maintain. ## What changes are included in this PR? - Use Arrow casts for Parquet microsecond-to-millisecond timestamp arrays and scalars, while preserving Spark's metadata-only NTZ-to-timezone relabeling. - Use Arrow's Date32-to-timestamp cast for the timezone-free path and retain the custom Spark DST handling for timezone-aware timestamps. - Use Arrow casts for millisecond-to-microsecond timestamps and Int32-to-Date32 arrays, and `ScalarValue::cast_to` for scalar Unix dates. - Strengthen coverage for non-UTC timezone relabeling, negative truncation, nulls, and millisecond-to-microsecond conversion. ## How are these changes tested? - `cargo test -p datafusion-comet-spark-expr --lib` (567 passed) - `cargo test -p datafusion-comet --lib parquet::cast_column::tests` (6 passed) - `cargo clippy -p datafusion-comet-spark-expr -p datafusion-comet --lib --tests -- -D warnings` - `cargo fmt --all -- --check` -- 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]
