Re: (Complex)Data Manipulation

[TedPride]

Very neat, though it assumes a rather precise format. My solution is more adaptive, if a good deal more messy:

@mon = ('','Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct'
+,'Nov','Dec');

$_ = <DATA>; chomp;
$n{$_} = $c++ for split /\t/;

for (<DATA>) {
    chomp; split /\t/;

    @date = split /\//, $_[$n{'Date'}];
    $product = $_[$n{'Product'}];
    $country = $_[$n{'Imported_From'}];
    $quantity = $_[$n{'Quantity'}];

    $countries{$country} = () if !exists $countries{$country};
    $totals{$date[2]}{$date[0]}{$product}{$country} += $quantity;
}

@countries = sort keys %countries;

for (sort {$a <=> $b} keys %totals) {
    $year = $totals{$_};
    for (sort {$a <=> $b} keys %$year) {
        $month = $year->{$_};
        print "$mon[$_]\n";
        print join("\t", 'Product', @countries, 'Total'), "\n";
        for (sort keys %$month) {
            print;
            $product = $month->{$_};
            $total = 0;
            for (@countries) {
                print "\t".$product->{$_};
                $total += $product->{$_};
            }
            print "\t$total\n";
        }
        print "\n";
    }
}

__DATA__
Date    Product    Quantity    Imported_From
1/2/04    Shirts    32    Australia
1/9/04    Shoes    234    Asia
2/12/04    Caps    109    UK
4/4/04    Shoes    6    Asia
4/4/04    Shirts    12    Australia
5/6/04    Shirts    398    Australia

You can have any number of fields in any order, and the records can also be out of order. You're not limited to three countries either. Results are given in tab delimited format, perhaps not the most pretty thing to use, but I didn't feel like messing with prints after spending so much time writing the algorithm.