adriangb commented on PR #22343: URL: https://github.com/apache/datafusion/pull/22343#issuecomment-4593888832
re dynamic vs. heuristic based: my feeling is we'll want dynamic adjustment long term. Neil is right that you need to "seed" in some way, and heuristic based is one option. But if it ends up being not much benefit and considerable code complexity it's not worth the step. If it ends up being very little code and helpful in reasonable cases then it's worth keeping. I do think there's also other heuristics we could use e.g. size of the columns the filter references from stats (so e.g. a like expression on a small column is not penalized as much as a large column, and could beat out a large mathematical expression). -- 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]
