Zouxxyy opened a new pull request, #7928:
URL: https://github.com/apache/paimon/pull/7928
### Purpose
Replace the ad-hoc column-alignment logic in `PaimonAnalysis` (a parallel,
top-level-only implementation of `TableOutputResolver`) with a dedicated
`PaimonOutputResolver` — a trimmed fork of `TableOutputResolver` with
`mergeSchemaEnabled` support that applies uniformly at every nesting depth.
- Missing target fields → NULL-fill at any depth
- Extra source fields → reject when `mergeSchemaEnabled=false`, else kept
inline so `SchemaHelper.mergeSchema` evolves the table at write time
- Nullable downcast → `AssertNotNull`; type mismatch → tagged `Cast`
`paimonWriteResolved` now mirrors Spark's V2 `outputResolved` strictness
(name + type + nullability, including nested). A `PAIMON_WRITE_RESOLVED`
`TreeNodeTag` prevents re-entry when extras are kept inline.
### Tests
Added recursive nested by-name/by-position alignment cases and merge-schema
conflict cases (deep nested array/struct, top-level same-name type mismatch)
in `DataFrameWriteTestBase`, `DDLTestBase`, `WriteMergeSchemaTest`.
--
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]
