dwsmith1983 commented on issue #25758: URL: https://github.com/apache/datafusion/issues/25758#issuecomment-5890993544
I'd like to propose #24790 for 55.2.0. It fixes #24786: when a file has an INT96 column, the INT96 coercion rebuilds struct, list and map fields without their metadata, so Parquet field ids on those containers are lost while leaf ids survive. Comet always enables the coercion, so with `spark.sql.parquet.fieldId.read.enabled` a container whose id sits only on the container is null filled, while Spark reads it (apache/datafusion-comet#6131). The fix adds `.with_metadata(...)` to the three container constructors in `schema_coercion.rs`, plus tests, and cherry-picks cleanly onto `branch-55`. I can open the backport PR. -- 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]
