dejankrak-db opened a new issue, #11876: URL: https://github.com/apache/arrow-rs/issues/11876
**Is your feature request related to a problem or challenge? Please describe what you are trying to do.** #10109 added the Parquet `FILE` logical type (schema and Thrift support). The Arrow schema mapping is still missing: a `FILE`-annotated group reads as a plain struct, and `ArrowWriter` cannot write the annotation, so Arrow-based writers cannot produce conforming `FILE` columns. **Describe the solution you'd like** Follow the Variant precedent in `parquet/src/arrow/schema/extension.rs`: - Read: tag a `FILE` group with a `file` extension type, `arrow.parquet.file`. - Write: emit `LogicalType::File` for a struct field carrying that extension type, validating its fields as #10109 does for the Parquet schema. - Gate it behind an experimental feature (`file_experimental`) until the extension name is settled. **Describe alternatives you've considered** Writers can build the Parquet schema themselves and pass it with `ArrowWriterOptions::with_parquet_schema`, but then every writer has to reimplement the mapping, and readers still get no indication that a struct is a `FILE`. **Additional context** `arrow.parquet.file` is the intended name for a canonical Arrow extension type, mirroring `arrow.parquet.variant`; it is not yet standardized. -- 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]
