aoto-tech commented on issue #25265: URL: https://github.com/apache/datafusion/issues/25265#issuecomment-5653473361
cosine_distance is particularly tricky. - array_distance: Relatively straightforward. Scale the differences by their maximum absolute value before computing the sum of squares. - array_normalize: Similar—divide by the largest absolute component before calculating the norm. - cosine_distance: More difficult because not only the norms but also the a * b terms in the dot product can overflow. The implementation also needs to handle cancellation between positive and negative values, zero vectors, and the semantics of NaN and Infinity. -- 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]
