I'm doing a data conversion project. I've written a lot of this kind of script. This first script reads in a comma-separated-value file, which contains lines which look like:

productid,long description goes here,acctnumber

It reads the file and creates two associative arrays for the second and third fields, keyed by the first field

#!/usr/bin/perl

use warnings;
use strict;
use vars qw/ %descs %salesAccts /;

open(UPRODFILE,'<uprod.csv');

while (<UPRODFILE>) {
    chomp;
    my $prodID = "";
    my $desc = "";
    my $salesAcct = "";
    ($prodID,$desc,$salesAcct) = split(/,/);
#    print "$prodID -- $desc -- $salesAcct \n";
    $descs{$prodID} = $desc;
    $salesAccts{$prodID} = $salesAcct;
}

close(UPPRODFILE);
#foreach my $i (sort keys %descs) {
#    print "$i == $descs{$i} == $salesAccts{$i} \n";
#}
1;
[download]

Then there's a second script that reads a second file, looks up each line in the two associative arrays, and creates a third file in the same format as the first.

#!/usr/bin/perl
require 'uprod.pl';

while (<>) {
    chomp;
    print "$_, $descs{$_}, $salesAccts{$_}\n"
}
[download]

This one runs in sloppy mode because I didn't want to generate warnings for cases where there was missing data. Usually I have warnings and strict turned on in every script - it just promotes better coding.

N.B. This requires that none of the fields in the first file contain embedded commas. If they do, this would have to be re-written using the Text::CSV module, which I have used on occasion, but avoid when unnecessary.

For me this is much faster and more predictable than using MS Excel with vlookup or something like that.