gabotechs commented on code in PR #25969: URL: https://github.com/apache/datafusion/pull/25969#discussion_r4181132075
########## docs/source/library-user-guide/upgrading/56.0.0.md: ########## @@ -428,6 +428,28 @@ part of `StatisticsRegistry::default_with_builtin_providers()`. - Remove `DefaultStatisticsProvider` from any custom provider chain; register no terminal provider instead (the walk falls back on its own). +### `StatisticsRegistry::default_with_builtin_providers` is deprecated + +`StatisticsRegistry::default_with_builtin_providers()` is deprecated in favor of +`StatisticsRegistry::with_experimental_providers()`. Most of the providers it Review Comment: For these providers, rather than just moving them to "experimental", it'd be nice to clarify their future and direction. For example, if they do are experimental, what's that scope of the experiment, is someone investing time in experimenting with them so that at some point they are eligible to replace their equivalent `ExecutionPlan::partition_statistics()` implementation? To be honest I don't really see much benefit in having two different concurrent implementations of the same statistics estimation code committed to `main` in order to keep one implementation as an "experiment". Typically experiments are something confined to a developer's PR until it proves an improvement via benchmarks, and once the improvement is qualified it replaces the previous code, rather than co-living as a mutually exclusive code path in `main`. -- 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]
