ritchie46 commented on pull request #9416: URL: https://github.com/apache/arrow/pull/9416#issuecomment-802691652
> Ok, there are some distinct views on the matter (mine included). I would assume users of the kernels also have varying requirements regarding NaN behavior. For this reason, I propose to leave the min, max kernels as is, and add additional min_float and max_float kernels in which users can define the NaN behavior: > Propogate: any NaN in array leads to NaN output. Ignore: (Postgres behavior) NaNs are ignore if any is aN. TotalOrder: Follow the total order standard. ? @alamb @jorgecarleitao @jhorstmann @Dandandan What do you think of letting the user define one of those predefined NaN behaviors? -- 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. For queries about this service, please contact Infrastructure at: [email protected]
