jackylee-ch opened a new pull request, #951: URL: https://github.com/apache/paimon-rust/pull/951
The Avro reader's `build_column` had no `VECTOR` arm and fell through to `Unsupported`, so a data file with a `file.format=avro` VECTOR column (e.g. an embeddings table) failed to read — even though Parquet reads the same column fine. Avro encodes a VECTOR identically to an ARRAY (an `array` of the element type), so this decodes it into an Arrow `FixedSizeList` of the declared length. A present row must carry exactly `length` elements (otherwise the data is rejected as malformed); an absent/null row contributes `length` masked placeholder slots to keep the fixed stride. Unit tests cover a `VECTOR<FLOAT, 3>` read with a null row, plus the wrong-length rejection. -- 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]
