MaxGekk commented on PR #55326: URL: https://github.com/apache/spark/pull/55326#issuecomment-4661803127
**PR description is out of sync with the diff (doc-only)** A few things in the description don't match the actual changes; worth trimming so reviewers aren't looking for code that isn't there: - "`ParquetWriteSupport` | Companion utility extraction (`consumeGroup`, `writeFields`)" - not in the diff. The only change in `ParquetWriteSupport` is the framework-first split of `makeWriter` into `makeWriter` + `makeWriterDefault`; `consumeGroup`/`writeFields` are untouched. - "`ParquetSchemaConverter` | ... + read-path reverse lookup" - the only change in the Parquet->Spark direction is a cosmetic line merge (`case _ => illegalType()`). No framework dispatch is added on the read/inference path; `TimeType` inference still uses the hardcoded `TimeLogicalTypeAnnotation` case. - The `ParquetTypeOps` scaladoc says the trait "covers all Parquet concerns: ... Vectorized read: creating batch updaters ... Filter pushdown: creating Parquet filter predicates". The trait actually has neither (they're correctly deferred to the follow-ups) - it only exposes the `isBatchReadSupported` gate. Suggest rewording the scaladoc and the description's trait-surface bullet list to reflect what's implemented now vs. deferred. None of these affect correctness; just aligning the narrative with the code. -- 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]
