[GitHub] [iceberg] RussellSpitzer commented on pull request #2891: Fix -NaN ordering in spec

[GitBox <gi...@apache.org>]

RussellSpitzer commented on PR #2891:
URL: https://github.com/apache/iceberg/pull/2891#issuecomment-1132242109

   I think i'm good with merging the patch as it was. According to IEEE we actually have a huge number of possible byte representations that are nan, 
   ```
   Any Sign bit : 2 Options
   All 1 for Exponent : 1 Option
   Anything but all 0 for the fraction : 2^23 - 1 options (or 2 ^ 53 -1) options
   ```
   so there are 2^24  -1 or 2 ^ 54 -1 possible NAN's for floats and doubles respectively.
   
   I don't think it's very important for us to define a concrete ordering of all different byte representations. If we want to be really strict we could just say all Nan's must be sorted in lexical byte order but I don't think it makes sense to put them on either end of the Infinity Spectrum. I think they should all just be grouped together regardless of if we strictly decide how they are ordered within that group.
   
   @rdblue Did you have strong opinions on this?


-- 
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: issues-unsubscribe@iceberg.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org