Setting $/ to undef (via local) makes it read the whole file as a single string, and not line by line. Which is why your approach fails.

My plan was to sort the file by the UID, and then check for lines with UID = to previous line. Is this the best way?

Not necessarily. The common perl idiom for checking uniqueness is to iterate over all the values, and store the UID in a hash. If it existed before in the hash, it's not uniq. Consider:

my %uids;
while (my $line = <$filehandle>) {
   my $uid = # extract from $line;
   if ($uids{$uid)) {
       print "'$uid' is duplicate\n"
   } else {
      $uids{$uid} = 1;
   }
}
[download]