etseidl commented on PR #221: URL: https://github.com/apache/parquet-format/pull/221#issuecomment-3142116207
> combined with the addition of two new statistics (`pos_nan_count` and `neg_nan_count`) I'm wondering if counts are actually necessary. I see a need for null counts, but with NaN I think all we really care about is the lack or presence of them, rather than the number. Perhaps two boolean fields `has_pos_nan` and `has_neg_nan`. Or maybe rather than adding two more fields we could instead add an `i32` or `i64` bitfield and an enum to define the bit positions. This would allow for other special cases in the future (say we want to exclude +/- infinity as well). > are you willing to to make such a proposal if it doesn't exist? I think #196 could be revived if need be. -- 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]
