voonhous commented on issue #18285: URL: https://github.com/apache/hudi/issues/18285#issuecomment-5931902019
The issues under this ticket split by schema path. Schema-on-read items are not backported to branch-1.2.x; schema-on-write bugs are. | Issue / item | Path | What it is | branch-1.2.x | |:--|:--|:--|:--| | #20135 | schema-on-write | A variant column added by `ADD COLUMNS` (v1 route) cannot be read from pre-DDL files with `pushVariantIntoScan` on | ported: #20184 | | #20139 | schema-on-read | `count(*)` on a variant table fails once an internal schema is committed | not backported | | #18551 | schema-on-read | Negative sentinel field ids for VARIANT and BLOB in `InternalSchemaConverter` | not backported | | #20141, schema-on-read part | schema-on-read | What blocks renaming a variant column (DDL, guard, codegen cast, catalog type) | not backported | | #20141, schema-on-write part | schema-on-write | No rename primitive: Spark refuses the DDL, a DataFrame write is drop plus add; its one read failure was #20135 | nothing to port | | #20140 | neither | The file format writes its batch decision into the session conf; not evolution-specific | separate decision | | Checklist items 1-3 | schema-on-write | Test-only pins (sibling widening on MOR and at top level, `ALTER COLUMN TYPE` refusal) | not needed | | Checklist item 4 | schema-on-write | Same as #20135 | ported: #20184 | | Checklist item 5 | schema-on-read | Test-only pin of the rename failure | not backported | The four operations in the scope comment above (add, drop, rename, reposition under `hoodie.schema.on.read.enable`) are all schema-on-read. -- 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]
