mkleen opened a new pull request, #26601: URL: https://github.com/apache/datafusion/pull/26601
## Which issue does this PR close? - Part of #26138. ## Rationale for this change DataFusion checks every float column used in a join, `GROUP BY` or comparison for `-0.0` values that need replacing. That check runs very often, and there's no benchmark for it yet. A follow-up PR for #26138 moves this code and makes the check faster. Adding the benchmark first means that PR can be compared directly against `main`. ## What changes are included in this PR? A new criterion benchmark, `float_canonicalization`, in `datafusion-common`. It measures `normalize_float_zero` on a batch of 8,192 values, for both `f64` and `f32`, in two cases: - `clean`: nothing to replace, the common case - `neg_zero_last`: one `-0.0` in the last row, so every value is checked and the column copied ## Are these changes tested? The benchmark runs: `cargo bench -p datafusion-common --bench float_canonicalization` ## Are there any user-facing changes? No. -- 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]
