ZhangHuiGui commented on PR #40998: URL: https://github.com/apache/arrow/pull/40998#issuecomment-2049117312
> > I've tested this, looks like a big performance improvement if use the resorted columns (perf shows it's benefit from easier vectorization). > > Thank you for both the test and the benchmark! Recently when I wanted to test the performance improvement of turning on the are_cols_in_encoding_order parameter in swiss-join, some unit tests failed. The RowTable participating in the comparison of Swiss-join has already been rearranged. In theory, turning on/off the are_cols_in_encoding_order parameter does not work. It will affect the correctness. It seems that we need to solve these problems of swiss_join, otherwise swiss-join will not be able to take advantage of the performance advantages after column-resort? BTY, the pr's ci-failure seems unrelated. -- 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]
