JingsongLi opened a new pull request, #8887:
URL: https://github.com/apache/paimon/pull/8887
## What changed
- Extract row-ID reassignment planning from `DataEvolutionRowIdReassigner`
into a focused planner.
- Add reusable projected `BinaryManifestEntry` and `BinaryDataFileMeta`
views.
- Bind projected fields by schema field ID instead of relying on a
planner-specific fixed layout.
- Keep the manifest scan hot path on binary strings and serialized rows
without materializing POJOs.
- Add dedicated unit tests for both binary views, including arbitrary
projections, unsupported fields, null handling, reuse, and lifecycle checks.
## Why
The row-ID reassignment implementation introduced in #8882 combines
orchestration, planning state, binary manifest decoding, and compact data
structures in one large class. This makes the algorithm difficult to review and
gives its inline binary decoding logic a schema that is specific to one caller.
The extracted binary views make projection handling reusable and keep the
planner focused on row-ID assignment.
## Impact
This is an internal refactor. Row-ID assignment behavior and manifest
serialization remain unchanged. Accessors for projected fields implement the
existing `ManifestEntry` and `DataFileMeta` contracts; access to fields that
were not projected fails explicitly.
## Validation
```text
mvn -pl paimon-core -DfailIfNoTests=false -DwildcardSuites=none \
-Dtest=BinaryDataFileMetaTest,BinaryManifestEntryTest,DataEvolutionRowIdReassignerTest
test
```
51 tests passed. Checkstyle, Spotless, and Enforcer passed.
--
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]
