Kontinuation opened a new pull request, #610: URL: https://github.com/apache/sedona-db/pull/610
## Summary - Inserts a `RepartitionExec` with `RoundRobinBatch(N)` partitioning as the parent of the probe side input to `SpatialJoinExec`, breaking spatial locality in sorted/skewed datasets to balance workloads during out-of-core spatial join. - Adds a new `repartition_probe_side` option to `SpatialJoinOptions` (default: `true`) to control this behavior. - Correctly handles non-KNN predicates with input swapping (`should_swap`) and KNN predicates where `probe_side` can be `Left` or `Right`. ## Changes - **`rust/sedona-common/src/option.rs`**: Added `repartition_probe_side: bool` (default `true`) to `SpatialJoinOptions`. - **`rust/sedona-spatial-join/src/planner/physical_planner.rs`**: Added repartition logic in `plan_extension()` and a new `repartition_probe_side()` helper function. The repartition is applied to the pre-swap input that will become the probe side after any potential swap. ## Testing All existing tests pass (332 unit tests, 169 integration tests, 3 doc-tests for `sedona-spatial-join`; 8 tests for `sedona-common`). Clippy clean. -- 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]
