JingsongLi opened a new pull request, #10605:
URL: https://github.com/apache/paimon/pull/10605
## Purpose
Use Rust core for the filesystem `$partitions` system table, which Python
already supports. Reuse `Table.new_system_read("partitions")` and the existing
in-memory system split; partition aggregation and deletion-vector counting
execute in Rust.
## Changes
- Enable Native partition metadata on the existing REST path.
- Align the Python implementation with Java: honor captured/time-travel
snapshots, preserve manifest-declared bucket counts, render typed and unescaped
partition specs with Java null/whitespace semantics, and sort output.
- Compare file creation times as instants, resolving DST gaps and overlaps
like Java, before rendering local TIMESTAMP metadata.
- Group using serialized partition values so identical display strings
cannot merge different partitions.
- Cover real Avro index-manifest cardinalities, unknown and absent counts,
occupied versus declared buckets, scalar partition casts, default names, and
captured splits. New Native tests are marked and skipped when the binding is
absent.
Depends on https://github.com/apache/paimon-rust/pull/1138. Keep this PR
draft until that change reaches Apache `paimon-rust@main`; CI continues to
build Native from `main`.
## Validation
- System-table and Native metadata regressions with Native plan/read enabled
against a locally compiled binding: 204 passed in UTC and 204 passed in
America/New_York, with no skipped tests.
- Existing `pypaimon/tests/py36`: 80 passed under the local Python 3.13
interpreter.
- Changed production files and the new test parse with Python 3.6 grammar.
- Changed-file Flake8: passed.
- Ordinary Python test selection excludes the new Native-only test module.
Independent review completed; Java whitespace handling and Native test
marking findings were fixed.
--
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]
