rdtr opened a new pull request, #13192:
URL: https://github.com/apache/gluten/pull/13192
## What changes are proposed in this pull request?
When the broadcast table is larger than `spark.sql.maxBroadcastTableSize`,
the Velox backend threw a plain `SparkException` with no error condition.
Vanilla Spark's `BroadcastExchangeExec` throws
`QueryExecutionErrors.cannotBroadcastTableOverMaxTableBytesError`, which has
the `_LEGACY_ERROR_TEMP_2249` condition. This PR throws the same error, through
a small helper in `BroadcastUtils` since `QueryExecutionErrors` is
`private[sql]`.
This is the `GlutenBroadcastExchangeSuite` item ("SPARK-56455: broadcast
should fail when table size exceeds maxBroadcastTableSize") in #13178. That
test is excluded for Spark 4.2 in #13163, so its exclude can be removed once
both are merged.
Part of #13178
## How was this patch tested?
Added a test to `VeloxHashJoinSuite` that checks the error condition, the
same check as the Spark test. It passes on Spark 4.1 with this change and fails
without it (the error condition is `null`).
## Was this patch authored or co-authored using generative AI tooling?
Generated-by: Claude Code (Claude Opus 5.5)
--
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]
