dbtsai opened a new pull request, #278:
URL: https://github.com/apache/spark-connect-rust/pull/278

   ### What changes were proposed in this pull request?
   
   Use each explicit pandas date column's exact NumPy backing array for both 
eligibility validation and Arrow conversion. Add regressions for customized 
`Series.values` and `Series.to_numpy`, and document the existing date importer 
path.
   
   ### Why are the changes needed?
   
   Follow-up to #138 / SPARK-60716. For 10,000 stored dates of `2020-01-01`, a 
customized `Series.values` returning `2021-02-03` made the fast path validate 
the stored cells but emit the replacement dates. This client's previous 
DataFrame row conversion retained `2020-01-01`. Validation and serialization 
now consume the same concrete storage.
   
   Canonical pandas column extraction through `iloc` and `Series.array` remains 
trusted, consistent with the existing string path; this change does not 
authenticate every mutable pandas descriptor.
   
   ### Does this PR introduce _any_ user-facing change?
   
   Yes. Customized Series accessors no longer replace the validated date cells 
or run during this optimization. Ordinary date values, null handling, names, 
thresholds and transport routing are preserved.
   
   ### How was this patch tested?
   
   - Both new accessor regressions fail against the landed helper; the updated 
helper passes all 26 selected helper cases (865 deselected).
   - Independent expected offsets and null masks pass for 30,000 cells spanning 
Python's full date calendar range; the regression checks Arrow IPC roundtrips 
and unchanged row normalization.
   - Matching official PySpark 4.3.0 conversion passes 10,000 ordinary 
full-range date/null values. The custom-accessor regression targets this 
client's previous row-path contract.
   - `git diff --check` passes. Genuine local Claude Code reviewed the full 
landed change and the complete follow-up patch; the final source review 
approved with no blockers and documented nonblocking limitations/nits.
   
   **Draft validation limitation:** helper checks use installed native schema 
types, not a fresh candidate extension. Cargo/Rust/protoc are unavailable on 
this host, so the required Cargo lib/all-integration/wasm compile/fmt/Clippy 
checks, fresh extension, native mock/live tests and full official parity/API 
gates remain pending. No overall validation convergence is claimed.
   
   ### Was this patch authored or co-authored using generative AI tooling?
   
   Generated-by: OpenAI Codex (GPT-6); reviewed with Claude Code 2.1.314.
   
   This pull request and its description were written by Isaac.
   


-- 
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