note sk [Tanktalus]'s suggestion is a better to handle this problem. But if you want something quick and dirty, you can do something like this - <p> <c> #!/usr/bin/perl -w use strict; my $base = "ca"; my @dirs = map { $base . int (rand 100) } (1..10); # Instead of matching for $base (ca) you can also match characters etc. This approach works for simple things and can get complex quickly my @sorted = sort { $a->[0] cmp $b->[0] || $a->[1] <=> $b->[1] } map { [/($base)(\d+)/] } @dirs; @sorted = map { $_->[0] . $_->[1] } @sorted; print $_, $/ for (@dirs); print $/; print $_, $/ for (@sorted); </c> <p> Output<p> <c> ca58 ca43 ca98 ca83 ca11 ca59 ca54 ca92 ca69 ca30 ca11 ca30 ca43 ca54 ca58 ca59 ca69 ca83 ca92 ca98 </c> 521694 521694