You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@iceberg.apache.org by GitBox <gi...@apache.org> on 2019/07/24 18:59:10 UTC

[GitHub] [incubator-iceberg] rdblue commented on issue #295: Normalize NaN values for bucketing

rdblue commented on issue #295: Normalize NaN values for bucketing
URL: https://github.com/apache/incubator-iceberg/issues/295#issuecomment-514758023
 
 
   I forgot that I had already thought about bucketing with floating point numbers and decided it was a bad idea. I'm glad you guys caught that in the spec (and that we wrote it down!).
   
   The reason to define a hash function for floating point numbers is that we may need it for other uses, and we need to be careful with hashing and schema evolution.
   
   The hash spec produces the same result for a value in all types the value may later be promoted to. That's why `h(1) == h(1L)`. That is required for bucketing to work when a bucketed column is promoted from `int` to `long`. While we don't support bucketing on floating point values, I think this property should be preserved if we ever do use the hash spec with floating point values.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services

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