If you are going to present us with error messages that include line numbers, you should indicate in your code which lines those numbers represent. I'm guessing that the two messages you quoted -- which are actually warnings (not errors) -- refer to the two lines marked below:

    my $one_number = $numbers[$hashtable{$name} - 1];  # line 24
    if ( $one_number >= 20 ) {
        print "$name $hashtable{$name} $one_number\n";   # line 26
    }
[download]

    next if ! s{ \A (\S+) \s+ (?= \d ) }{}xms;
[download]

    next unless ( s{ \A (\S+) \s+ (?= \d ) }{}xms and exists( $hashtab
+le{$1} ));
[download]

#!/usr/bin/perl -w

  my %hashtable;

      my $fn = <>;  ##### File 1 is the input.

      open(FH, "$fn") || die "Cannot open file";
[download]

First, do a favour to yourself and use strict as well. Then, if you don't chomp $fn, it will end with "\n" and be unlikely albeit not impossible to be the real filename you're after. Last, $fn is called "useless use of quoting." Well, it's not really "last" because I may point out that you'd better use lexical filehandles, the three args form of open, low precedence short circuiting logical operators, include both the name of the file and more importantly the cause of the error ($!) in the error message, etc. But I'll try to stick to pointing out only the major issues...

    my $one_number = $numbers[$hashtable{$name} - 1]; ### Error warnin
+g in this line

    if ( $one_number >= 20 ) {

        print "$name $hashtable{$name} $one_number\n";  ### Error warn
+ing in this line
[download]

Indeed: what if $name is not a key of $hashtable? Appearently you have one or more names in the second file that are missing from the former. Try finding out which ones they are. A simple

say "<$name>" unless exists $hashtable{$name};
[download]

would do: the additional < and > marks are there to be sure about "hidden" characters like spaces. But if you want to be really fussy you may even try:

say map {my $_=$_; s/([^[:print:]])/sprintf '\\%03o', $1/ge; "<$_>"} $
+name
  unless exists $hashtable{$name};
[download]

I don't understand why they used <>. This returns the first line of the first file passed. Not the first argument on the command line.
You're right they'd have to chomp it if the filenames were listed at the start of each file but seeing as I don't think they meant that. shift of the @ARGV array is fine and wouldn't need chomping.

I personally believe that he's not passing any file on the command line, (the other filename is hardcoded some lines below...) in which case it will read the first line from STDIN, but certainly, if he wanted to rely on this behaviour then he should have used <STDIN>. Just one out of many gotchas in a lump of bad code.

--
If you can't understand the incipit, then please check the IPB Campaign.

#!/usr/bin/perl
use warnings;
use English q/-no_match_vars/;
use strict;
use Carp qw/croak carp/;
use Data::Dumper;

my $file_one_ref;
my $file_one_fn = shift or croak "No input file specifed for file one"
+;
my $file_two_fn = shift or croak "No input file specifed for file two"
+;

open my $file_one_fh, '<', $file_one_fn
  or croak "Could not open filename [$file_one_fn] : ", $OS_ERROR;

while (<$file_one_fh>) {
    $file_one_ref->{ lc $1 } = $2 if / \A (\w+) \s+ (\w+) /xms;
}

# You can do the above input like this but it's not as efficient:
# map { $file_one_ref->{ lc $1 } = $2 if / (\w+) \s+ (\w+)/xms; } <$in
+put_fh>;

# Have a look at the keys read in:
print Dumper($file_one_ref), "\n\n";

open my $file_two_fh, '<', $file_two_fn
  or croak "Could not open filename [$file_two_fn] : ", $OS_ERROR;

$INPUT_RECORD_SEPARATOR = '>';

while ( my $line = <$file_two_fh> ) {

    # I reserve s{}{}xms; for multi line expressions
    next if ( $line !~ s/ \A (\S+) \s+ (?= \d ) //xms and !exists $fil
+e_one_ref->{ lc $1 } );

    # why not just use a match? is $name going to contain numbers

    my $name = lc $1;
    my $numbers = [ grep { $_ } split /\D+/, $line ];
    my $one_number = $numbers->[ $file_one_ref->{$name} - 1 ] || next;
    if ( $one_number >= 20 ) {
        print join " ", $name, $file_one_ref->{$name}, $one_number, "\
+n";
    }
}

close $file_one_fh
  or croak "Could not open filename [$file_one_fn] : ", $OS_ERROR;
close $file_two_fh
  or croak "Could not open filename [$file_two_fn] : ", $OS_ERROR;

__END__
[download]

If you're going to use English and aren't actually using the verbose forms of $& and friends it's better to use English '-no_match_vars'; instead so as not to slow down every regex match your program does (presuming a post 5.8 perl; before that there's not a way to dodge the slowdown bullet).

(But really, bite the bullet and learn which is which; or just remember: perldoc perlvar is my friend.)

The cake is a lie.
The cake is a lie.
The cake is a lie.

Thanks, Always learning.
John