zjw1111 opened a new issue, #296: URL: https://github.com/apache/paimon-cpp/issues/296
### Problem Paimon tables written with Arrow-based Parquet writers can carry `ARROW:schema` metadata whose logical Arrow type differs from the physical type expected by the Paimon reader. For scalar BLOB columns, PyArrow may restore a physical Parquet `binary` column as Arrow `large_binary` before Paimon applies its binary-to-BLOB conversion. The strict type check then rejects compatible data. For nested `MAP<STRING, BLOB>` values, Rust Paimon writes Arrow `binary`, while the logical Paimon read type is `large_binary`; this currently fails during nested Parquet projection. ### Expected behavior The Parquet reader should distinguish representation-compatible fields from fields requiring conversion, preserve existing timestamp adaptation, and support both scalar and map BLOB compatibility without relaxing unrelated type validation. ### Test coverage Add Python, Rust, and Java append-table fixtures that validate all compatible values, read `MAP<STRING, BLOB>` successfully, and continue to report an explicit unsupported error for `ARRAY<BLOB>`. -- 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]
