parthchandra opened a new pull request, #3915:
URL: https://github.com/apache/datafusion-comet/pull/3915
## Which issue does this PR close?
Running TPC-DS resulted in a schema mismatch error in q36a-v2.7
## Rationale for this change
The golden file for TPC-DS v2.7 query q36a was originally generated against
Spark 3.4, where DecimalPrecisionTypeCoercion.bounded() simply clamped
precision/scale to the maximum (min(p, 38), min(s, 38)). This caused the UNION
of decimal(37,20) and decimal(38,11) (the two branches of results_rollup) to
widen to decimal(38,20).
[SPARK-45905](https://issues.apache.org/jira/browse/SPARK-45905) changed the
decimal least-common-type algorithm to prefer retaining integral digits over
fractional digits when precision exceeds 38. The same UNION now correctly
widens to decimal(38,11). Apache Spark's own TPCDSQueryTestSuite golden file
reflects this.
## What changes are included in this PR?
This regenerates the Comet golden file against Spark 4 to match the updated
behavior. Also, decimal tests now also check that the schema matches with the
one output by Spark.
## How are these changes tested?
Running TPC-DS queries in `CometTPCDSQueryBenchmark`. Also unit tests.
--
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]
