Question

I smell something bad here?

    if ($col == 24) {
        $buffer{'Y'} = trim($val);
        return;
    }

    if ($col == 25) {
        $buffer{'Z'} = trim($val);
        return;
    }

    if ($col == 26) {
        $buffer{'AA'} = trim($val);
        return;
    }

    if ($col == 27) {
        $buffer{'AB'} = trim($val);
        return;
    }

Answer 1

The names look suspicious, too. If you’re looping through the columns, try the magic ++ operator.

my $colname = 'A';
for (0..$#cols)
{
   # do stuff with $colname
   $buffer{$colname} = trim($val);
   ++$colname;
}

If not, there does seem to be a pattern here that you can exploit for converting numbers from decimal (digital) to alphabetic. You’d do it the same way you’d convert a digital number to decimal, except that you’d use characters A-Z, base 26, instead of 0-9, base 10. Something like:

sub colname
{
  my $num = shift;
  my $name = '';
  while ($num)
  {
    $name .= chr(ord('A') + $num % 26);
    $num /= 26;
  }
  reverse $name;
}

(untested) This algorithm is language-neutral. It doesn’t particularly take advantage of perlishness, but works wonderfully as a general-case.

Update: I told you this was untested. j_random_hacker pointed out the thinko, and I’ve corrected it. Thanks!