Hello htmanning,

Building on NetWallah’s answer: Since you say the URLs in question are “long,” it might be unsafe to assume that the ####.htm pattern (where # is a digit) never appears in the earlier part of a URL. In which case, play safe and anchor the match to the end:

#! perl
use strict;
use warnings;

my $re = qr{ ( \d{4} ) \. htm l? $ }ix;

/$re/ && print "$1\n" while <DATA>;

__DATA__
http://test-text-ad-1234.html
http://test-abc0000.htmdef-ad-5678.HTM
test-ghijk-ad-XXXX.html
[download]

Output:

12:39 >perl 1616_SoPW.pl
1234
5678

12:39 >
[download]

Hope that helps,