Re: My first perl script is working, what did I do wrong?

[rjt]

You have a lot of options in your approach, here. I humbly offer my somewhat more Perl-ish approach to this problem.

The basis of what I did is to get rid of hard-coded fields and allow you to specify them near the top of the script in @fields, building a regexp to capture them in the header line. I relaxed the usage of the program as well, to allow for specifying a filename on the command line, or, if none is specified, accepting input from STDIN. You can easily change the output format by modifying the final for (sort keys %inv) { ... } loop, and you can change how records are grouped by changing the key() function to suit your tastes. In both cases, I tried to stay with what you had, since I don't know what you want to ultimately do with the data.

use strict;
use warnings;
use 5.12.00;

die "usage: $0 [filename]\n" if (@ARGV > 1);

# Set required fields and determine fixed widths
my @fields = qw/Part Shape Color Size/;
my $re;
$re .= qr/(?<$_>$_\s*)/ for @fields;

my $header = <>;
die "Header must match " . join(' ', @fields) unless ($header =~ /^$re
+$/);
my $expected_len = length($header);

my $tmpl = join(' ', map { "A[".length($+{$_})."]" } @fields);

my %inv; # Inventory; $inv{key(%rec)}

while (<>) {
    if (length != $expected_len) {
        die sprintf("Got length of %d, expecting %d", length, $expecte
+d_len);
    }
    my @rec = map { /(.+?)\s*$/ } unpack $tmpl; # Get (trimmed) record
+s
    my %rec = map { $_ => shift @rec } @fields;

    # Add the part to our inventory
    push @{$inv{key(%rec)}}, $rec{Part};
}

# Now print out the inventory in the desired format
for (sort keys %inv) {
    say join(' ',@{$inv{$_}}) . " - $_ - " . scalar @{$inv{$_}}; 
}

# Our custom hashing function for inventory items. Expects %rec argume
+nt
sub key { my %rec = @_; $rec{Shape} . $rec{Color} . $rec{Size}; }

[killersquirel11]

Thanks for the reply, it actually looks like Perl! A couple quick questions on this:
1) Are both 'use 5.12.00' and 'use strict' necessary? From what I read, 'use 5.12.00' implies 'use strict'.
2) With the following line, I'm having a bit of difficulty understanding it.

$re .= qr/(?<$_>$_\s*)/ for @fields;

Is it essentially this?:

 
for my $field (@fields) $re=$re.qr/(?<$_>$_\s*)/

[rjt]

1. You are correct, use strict is not required here. Sorry, old habits die hard. :-)
2. Not quite. Your example will not compile, but a similarly expanded translation of what I wrote would look like:

        for my $field (@fields) {
            $re .= qr/(?<$field>$field\s*)/;
        }

Read $_ for a description of the $_ special variable, but the short version is, if you do not supply a variable name to for, Perl will automatically assume $_.

The regex itself might benefit from a bit more explanation. It's capturing all of the column names including trailing whitespace, and saving those as named captures in %+ for use in the line that builds $tmpl:

    my $tmpl = join(' ', map { "A[".length($+{$_})."]" } @fields);

That generates an (un)pack template string based on the field lengths read from the header, determining the length of each field in @fields by taking the length of the same-named capture in %+ . I'm using map here to "map" the values in @fields to the list that I want: a list of the lengths of each field.