alamb commented on PR #22712: URL: https://github.com/apache/datafusion/pull/22712#issuecomment-4672476038
> Summary: med/high cardinality is faster; low cardinality can be slower but acceptable?; high cardinality is slower due to a missing fast path, see below. Another thing I have been thinking about for the low cardinality case would be some sort of dynamic cut over -- for example, maybe after your refactoring we can use a direct group implementation when there are less than batch_size groups, and once we switch to more than batch_size groups, switch to a two part key 🤔 -- 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]
