andygrove opened a new issue, #3316: URL: https://github.com/apache/datafusion-comet/issues/3316
## Summary 5 Spark SQL tests fail because `native_datafusion` doesn't respect Parquet field ID matching. ## Failing Tests - `ParquetFieldIdIOSuite`: "Parquet reads infer fields using field ids correctly" - `ParquetFieldIdIOSuite`: "absence of field ids" - `ParquetFieldIdIOSuite`: "SPARK-38094: absence of field ids: reading nested schema" - `ParquetFieldIdIOSuite`: "multiple id matches" - `ParquetFieldIdIOSuite`: "read parquet file without ids" - `ParquetFieldIdIOSuite`: "global read/write flag should work correctly" ## Root Cause `native_datafusion` reads columns by name/position rather than Parquet field IDs, producing wrong results when `spark.sql.parquet.fieldId.read.enabled` is true. ## Possible Fix In `CometScanRule.nativeDataFusionScan()`, detect when `spark.sql.parquet.fieldId.read.enabled` is true and fall back to `native_iceberg_compat`. ## Related Discovered in CI for #3307 (enable native_datafusion in auto scan mode). -- 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]
