wirybeaver commented on PR #3342: URL: https://github.com/apache/iceberg-rust/pull/3342#issuecomment-6621274931
@CTTY @paleolimbot Thanks for clarifying the distinction between computing bounds and propagating existing statistics. I've added a separate commit restoring the use of supplied Parquet bounds for both Geometry and Geography. This path consumes `ColumnChunkMetaData::geo_statistics()`, merges the row-group boxes, and serializes Iceberg file bounds. Computing bounds that include curved edges remains the Parquet producer's responsibility; this code does not derive bounds from WKB vertices. Geography boxes preserve antimeridian wrapping, with full longitude coverage represented as `[-180, 180]`. If any non-null row group has no box, the file bounds remain unknown. With Parquet's default Geography accumulator, which provides no bounds, they are still omitted; bounds supplied by a custom accumulator can now be preserved. The other requested changes remain in place: the field type travels with the accumulator state, and the tests are split by purpose with explicit XY, XYZ, and wrapping-Geometry coverage. Added focused regressions for supplied Geography boxes and full-longitude unions. All 34 writer tests, `cargo check`, formatting, and Clippy checks pass. -- 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]
