[spark] Column projection broken on log/upsert read path for type-dependent accessors

[fresh-borzoni]

Search before asking

• I searched in the issues and found nothing similar.

Fluss version

0.9.0 (latest release)

Please describe the bug 🐞

FlussPartitionReader.rowType uses full table schema, but scan returns projected rows. FlussAsSparkRow dispatches getLong, getBinary, getStruct, getArray, and copy based on rowType.getTypeAt(ordinal) - when projection shifts ordinals, the wrong type is looked up. Affects FlussAppendPartitionReader and FlussUpsertPartitionReader.

Solution

Override convertToSparkRow row type in subclasses - project(projection) for append, project(projectionWithPks) for upsert.

Are you willing to submit a PR?

• I'm willing to submit a PR!