andygrove opened a new issue, #2660:
URL: https://github.com/apache/datafusion-comet/issues/2660
### Describe the bug
When using `native_datafusion` scan, we sometimes try and construct a
`HashJoinExec` with different number of partitions for left and right inputs.
This issue is currently hidden because we wrap the inputs in a `CopyExec` which
always reports output partition count of 1.
Here is debug logging for one example. The left input has 5 partitions and
the right input has 1 partition.
```
LEFT: FilterExec: c0@0 IS NOT NULL
DataSourceExec: file_groups={5 groups:
[[tmp/CometFuzzTestSuite_1761747873664.parquet/part-00002-e5c142ac-d0f8-4eb4-891c-4484865ded05-c000.snappy.parquet:0..133654],
[tmp/CometFuzzTestSuite_1761747873664.parquet/part-00003-e5c142ac-d0f8-4eb4-891c-4484865ded05-c000.snappy.parquet:0..133566],
[tmp/CometFuzzTestSuite_1761747873664.parquet/part-00000-e5c142ac-d0f8-4eb4-891c-4484865ded05-c000.snappy.parquet:0..133492],
[tmp/CometFuzzTestSuite_1761747873664.parquet/part-00004-e5c142ac-d0f8-4eb4-891c-4484865ded05-c000.snappy.parquet:0..133471],
[tmp/CometFuzzTestSuite_1761747873664.parquet/part-00001-e5c142ac-d0f8-4eb4-891c-4484865ded05-c000.snappy.parquet:0..132287]]},
projection=[c0], file_type=parquet, predicate=c0@0 IS NOT NULL,
pruning_predicate=c0_null_count@1 != row_count@0, required_guarantees=[]
RIGHT: ScanExec: source=[BroadcastQueryStage (unknown),
Statistics(sizeInBytes=1384.0 B, rowCount=1.00E+3)], schema=[col_0: Boolean]
```
### Steps to reproduce
_No response_
### Expected behavior
_No response_
### Additional context
_No response_
--
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]
