hhhizzz commented on PR #10141: URL: https://github.com/apache/arrow-rs/pull/10141#issuecomment-5018312161
#10288 is about to land and changes Mask execution across page-pruned data. When rebasing this PR, could we preserve its `LoadedRowRanges`-aware Mask path and thread `loaded_row_ranges` into both cursor constructors, especially: ```rust new_mask_from_buffer(mask, loaded_row_ranges) ``` #10288 only has `new_mask_from_selectors`, while this PR introduces the direct `BooleanBuffer -> MaskCursor` path. If the ranges are not propagated there, `RowSelection::from_boolean_buffer(...)` can still cross an unloaded page and reproduce the sparse-page offset failure. Could we also add a sparse-page regression test starting from a direct, preferably non-byte-aligned, `BooleanBuffer::slice(...)`? The existing #10288 tests all start from `Vec<RowSelector>`, so they do not cover this integration seam. The old `should_force_selectors` / `override_selector_strategy_if_needed` fallback should be removed after the rebase, since #10288 makes Auto remain Mask safely across skipped pages. -- 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]
