thinker's solution is perfectly valid. However it's probably not the most efficient way to do it. This may be a concern if you have a lot of data.

So, I'll suggest using a Schwartzian Transform (an algorithm created by our very own merlyn).

untested code:

#!/usr/bin/perl -w

use strict;

open (DATA, "./data.txt") or die $!;  # for example

my @sorted = map { $_->[1] }
             sort { $a->[0] cmp $b->[0] }
             map { [ &fix_date($_), $_ ] } <DATA>;

for (@sorted) { print $_ };

sub fix_date {
   my ($d,$m,$y) = ( shift =~ m|(\d\d)/(\d\d)/(\d{4})| );
   return "$y-$m-$d";
}
[download]

The basic speed eater for thinker's way is that you have to munge the date to sort it a great number of times.

The first (lower) map does the time expensive data munging, into a form that's easy for sort to deal with, and the last (top) map puts the data back into the original form

HTH

/\/\averick
perl -l -e "eval pack('h*','072796e6470272f2c5f2c5166756279636b672');"