Perhaps if you give us a few lines of input and tell us what is supposed to match, and if so, what part of that string you expect your regexp to pull out, and we can try and point you in the right direction (or even write it).

Ok, here's (hopefully) a more complete example: Example strings to match:

confederation-as-router: 1 2345 atlanta1-cr1
confederation-as-router: 1 2345 atlanta1-cr2
confederation-as-router: 1 2345 atlanta1-cr3
confederation-as-router: 1 2345 atlanta1-cr4
confederation-as-router: 1 2345 atlanta1-cr5
confederation-as-router: 1 2345 atlanta1-cr6
confederation-as-router: 1 2345 atlanta1-sr1
[download]

open (FILE2, $file_2) or die "Couldn't open file: $!"; while (<FILE2>) { if ($_ =~ s/^confederation-as-router\:\W+\d\W+\d+\W//) { push(@file2, $_); } } </CODE> Currently, the regex will return:

atlanta1-cr1
atlanta1-cr2
atlanta1-cr3
atlanta1-cr4
atlanta1-cr5
atlanta1-cr6
atlanta1-sr1
[download]

Based upon the above example, I want it to ONLY return:

atlanta1-sr1
[download]

Is that clearer?

It might be simplest to do this:

print if !/cr\d$/ && $_ = (split)[3];
[download]

You're basically breaking apart by spaces anyway, and if you can assume that each line will follow this pattern, it might be simplest to use split and only use a regular expression to be sure you're working with the right line. Even if you do use a regular expression like yours, it's probably easiest to break it apart anyway:

if (!/cr\d$/ && s/^confederation-as-router:\W+\d\W+\d\W//) {
   ...
}
[download]

or even

if (!/cr\d$/ && ($host, $last) = (split)[0,3]) {
   print $last if $host eq 'confederation-as-router:';
}
[download]

Don't do that with a single regex.

while( <FILE> ) {
    if(  s/^confederation-as-router\:\W+\d\W+\d+\W//
     &&  ! /cr\d$/  ) {
        push(@file2, $_);
    }
}
[download]

        - tye (but my friends call me "Tye")