I noticed a speed improvement by changing LC_ALL from “C.UTF-8” to “C” from 22s to 1.3s. This is huge, and should not be ignored.
My setting: I call sort in a Dockerfile, derived from the official “python” base image. It sets LANG to “C.UTF-8”. Thus, my setting is not exotic. While I don’t really understand why UTF-8 encoding has that much impact on sorting performance, it may well be. However, this should be mentioned in the documentation in my option. Something like “Note that anything but plain C local may have significant impact on sorting performance” should occur somewhere in the man and info pages. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/846628 Title: gnu sort extremely slow in non C locale To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/coreutils/+bug/846628/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs