Hello, > On Jul 11, 2016, at 21:38, David Pan <nanosp...@139.com> wrote: > > It seems that I find a bug when using the command: > > sort -k1.4n -u > please refer to the attachment for detail .
You have not indicated exactly what is the suspected incorrect output. May I ask you to detail what is the output you expected vs the output you got? (i.e. in which command you think the bug is?) Also, in the subject line you've listed "-k1.4,1.6n", but I do not see such invocation in the attached file. To help troubleshoot sort-related issues, recent versions of gnu sort support a new option "--debug" which prints additional information about the keys being compared. It would be useful to attach the output of such command, for example: ====== $ sort --debug -k1.4n -u aa sort: using ‘en_US.UTF-8’ sorting rules sort: leading blanks are significant in key 1; consider also specifying 'b' sort: key 1 is numeric and spans multiple fields c3a1.ecld.com __ c3m2.ecld.com __ c3s15.ecld.com ___ c3s16.ecld.com ___ c3s17.ecld.com ___ c3s18.ecld.com ___ c3s19.ecld.com ___ c3s20.ecld.com ___ c3s21.ecld.com ___ c3s25.ecld.com ___ ====== regards, -assaf P.S. If you do test a newer version of coreutils, please be sure to use version 8.19 or newer, as it contains a fix for a "sort -u" bug which was introduced in 8.6 ( http://git.savannah.gnu.org/cgit/coreutils.git/commit/?id=eb3f5b3b3de8c6ca005a701f09bff43d778aece7 ).