andygrove opened a new issue, #3306:
URL: https://github.com/apache/datafusion-comet/issues/3306

   ## Description
   
   `CometScanRule` always uses `native_comet` for V2 scan handling, even when a 
different scan implementation is configured via `spark.comet.scan.impl`. If the 
user has configured `native_datafusion` or `native_iceberg_compat`, the V2 path 
silently ignores that setting and uses `native_comet` instead.
   
   ## Expected Behavior
   
   If the user has explicitly configured a scan impl other than `native_comet`, 
and V2 scans only support `native_comet`, the rule should fall back to Spark 
rather than silently using `native_comet`.
   
   ## Context
   
   This was discovered while running Spark 3.5.7 SQL tests with 
`native_datafusion`. The `ParquetV2FilterSuite` and `ParquetV2QuerySuite` tests 
passed, but only because V2 scans were not actually using the configured 
`native_datafusion` scan — they were silently falling back to `native_comet`.


-- 
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]

Reply via email to