Pandas886 opened a new pull request, #508: URL: https://github.com/apache/paimon-rust/pull/508
### Purpose Linked issue: close #503 I'd like a small public API for batch incremental reads — basically "files between snapshot A (exclusive) and B (inclusive)" — without callers reinventing snapshot walking on top of `TableScan`. This PR is intentionally thin: **Delta + Auto selection only**. Changelog / Diff come in follow-up PRs so review stays focused. ### Brief change log - Add `IncrementalScanMode`, `IncrementalScan`, `IncrementalPlan`, `IncrementalSplit` - Range semantics: left-open / right-closed, empty range, out-of-bounds errors - Delta: plan data files from APPEND snapshots in the range via existing scan helpers - Auto: `changelog-producer=none` → Delta; otherwise resolves to Changelog (planning Changelog still fails loud until the next PR) - Push partition/bucket filters through from `ReadBuilder` when planning - `TableRead::to_incremental_arrow` for the data-split case No new manifest reader code — it reuses what `TableScan` already does. ### Tests ```bash cargo test -p paimon --test incremental_batch_scan_test cargo test -p paimon --lib cargo fmt --all -- --check cargo clippy -p paimon --all-targets --features fulltext,vortex -- -D warnings ``` ### API and Format - New public scan types under `paimon::table` (also re-exported at crate root) - No storage format change ### Documentation None yet — waiting until the SQL TVF lands so we document the whole story once. Can add a short rustdoc note if preferred. ### Follow-ups - #504 Changelog + AuditLogTable - #505 Diff mode - #506 DataFusion `paimon_incremental_query` -- 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]
