tdunning commented on pull request #2432: URL: https://github.com/apache/drill/pull/2432#issuecomment-1022642104
If you want equal absolute error across the range then you can set the scale function on the t-digest to K_0 so it doesn't prioritize the tails. This can give roughly 2:1 improvement on median accuracy. On Wed, Jan 26, 2022 at 2:27 AM James Turton ***@***.***> wrote: > @tdunning <https://github.com/tdunning>, that would be great thanks. Here > the selectivity of a range predicate on an integer column is estimated from > an a set of quantile estimates, here termed an equidepth histogram (see > NumericEquiDepthHistogram > <https://github.com/apache/drill/blob/4aefcef2b665c5737471664912a26ef6ed9a6cfc/exec/java-exec/src/main/java/org/apache/drill/exec/planner/common/NumericEquiDepthHistogram.java>). > The quantile estimates are recovered from a MergingDigest over the integer > column that was created by an earlier ANALYSE TABLE command. If said > quantile estimates have shifted then that would explain what we're seeing... > > — > Reply to this email directly, view it on GitHub > <https://github.com/apache/drill/pull/2432#issuecomment-1022065612>, or > unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAB5E6WPTDXIWKPQKL6TAK3UX7EBPANCNFSM5MM2LDFQ> > . > Triage notifications on the go with GitHub Mobile for iOS > <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> > or Android > <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. > > You are receiving this because you were mentioned.Message ID: > ***@***.***> > -- 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]
