haohuaijin opened a new issue, #26030: URL: https://github.com/apache/datafusion/issues/26030
### Is your feature request related to a problem or challenge? PR #25731 preserves bitmap-backed Parquet row selections. When intersecting an existing bitmap-backed selection with a selector-backed selection, `ParquetAccessPlan::scan_and_select` currently builds a bitmap manually using `BooleanBufferBuilder`. The upstream API proposed in [apache/arrow-rs#11337](https://github.com/apache/arrow-rs/pull/11337), `RowSelection::force_mask()`, encapsulates this conversion. This issue tracks adopting it once DataFusion depends on a released parquet version that includes the API. ### Describe the solution you'd like After upgrading to a parquet version that provides `RowSelection::force_mask()`: - Replace the manual selector-to-bitmap conversion in `datafusion/datasource-parquet/src/access_plan.rs` with `selection.force_mask()`. - Remove any imports made unused by the change. - Verify that the existing row-selection intersection tests continue to pass, preserving bitmap backing and selected rows. ### Describe alternatives you've considered Keep the current manual conversion. It preserves the required behavior, but duplicates functionality provided by parquet once the new API is available. ### Additional context - Follow-up to [the review discussion in #25731](https://github.com/apache/datafusion/pull/25731#discussion_r4164530533). - Upstream API: [apache/arrow-rs#11337](https://github.com/apache/arrow-rs/pull/11337). -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
