Abhisheklearn12 commented on PR #25771: URL: https://github.com/apache/datafusion/pull/25771#issuecomment-5957740559
> @Abhisheklearn12 I'm not seeing any improvementins in the benchmarks and `short_circuit/and/all_false` seems to have regressed - is this expected? tbh, i partly expected this: - **no improvements:** very sorry, my bad: i should have merged `main` after #25870. this branch was cut before it merged, so neither side had the `conjunction` benchmarks, and the existing `short_circuit` cases aren't the ones this pr targets. in the same round, tpc-h q6 and q12 were 1.18x and 1.17x faster. i've merged `main` now, so the next `binary_op` run should include them. - **`all_false`:** yes, that's a real regression. when the first conjunct already rules out every row, `main` returns right away, while this pr first flattens the chain of conjuncts before it can stop (about 40 to 50 ns per batch). i'll remove that and post numbers. -- 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]
