JingsongLi opened a new pull request, #1484:
URL: https://github.com/apache/paimon-rust/pull/1484

   ### Purpose
   
   Match apache/paimon#11326. Two finite scores of opposite sign can overflow 
the span used by `weighted_score`. This can return the wrong Top-K for Data 
Evolution or cause primary-key ranking to reject a NaN score.
   
   ### Brief change log
   
   Share the route normalization between Data Evolution and primary-key 
ranking. Use double precision for the overflowing span and retain the 
single-precision path for ordinary scores.
   
   ### Tests
   
   Both new regression tests failed on the baseline. The 47 hybrid-search and 
primary-key ranker tests passed, and `cargo fmt --all --check` passed.
   
   ### API and Format
   
   No API or storage format changes.
   
   ### Documentation
   
   The documented min-max normalization now also holds for finite inputs whose 
span exceeds the single-precision range.
   


-- 
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]

Reply via email to