andygrove commented on PR #6447: URL: https://github.com/apache/datafusion-comet/pull/6447#issuecomment-5981351428
Addressed in a0b1f4db5: - `float_comparisons.sql` runs `IS DISTINCT FROM` end to end, on scalars, constants, arrays and structs. - A `float_cmp_deep` table compares `ARRAY<STRUCT<v: DOUBLE>>` and `STRUCT<a: ARRAY<DOUBLE>>` columns under every ordering operator, and negates the leaves so that a NaN two levels down has the sign bit set. Stopping `normalize_nested_floats` after one level fails only those queries. - The pruning test now puts a NaN in a row group whose other values are all below 500, and checks that `d > 500.0D` still returns it. parquet-mr writes the NaN as the chunk's max statistic, so DataFusion keeps that row group. - Agreed on the copies for nested ordering. The fix is a comparator that orders nested floats the Spark way in place, as nested `=` and `<>` already compare; that follow-up is #6590. -- 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]
