Rebecca Schlussel

thanks for review @abhiseksaikia and @ClarenceThreepwood. I've addressed your comments. I also split out the commits a bit as per request from @ajaygeorge.

Absolutely!

Hi @bearownage. I believe this issue was actually fixed by https://github.com/prestodb/presto/pull/19465. I'll close the issue.

I see the logic behind this change, as it seems to achieve what the original function was trying to do. However, I agree with @spershin that we should wait to...

That's fair. Yes, I agree the old behavior is very clearly a bug. And given that someone has taken the time to fix it, you're right, we should merge it....

yes, let's close this one

I've been looking into the NaN inconsistencies, and wanted to give an update/summary of my thinking. If we can get alignment on the solution we can start tackling the problem...

> @rschlussel Not sure what users we have in mind but from a numerical computing background, treating NaN as a normal number is very counter intuitive (making it larger than...

> @czentgr proposed the following behavior for Velox: [facebookincubator/velox#7237](https://github.com/facebookincubator/velox/pull/7237) Thank you! It sounds like we are in agreement then. Let me collect approvals so I can move forward with this...

Thank you, yes, i did know about this difference. I made a wiki about how we handle nans across all of our functions that do comparison or ordering on doubles...