I think an AoA should be what you need. Something in the lines of:

my $beans = [ [ 1 , 2 ], [2 , 3] ];
[download]

Here, $beans->[0][0] would have 1 bean, $beans->[0][1] 2, etc...
You don't need a "3rd dimension" for the array beacause you can have the number of beans as value.

Hope it helps,
deibyz

Note: Code untested.

1. A 3D array, as you say, which would actually be an array of array references of array references. A possibility, but probably unnecessary, as your indices would be x,y,beans - and that would probably waste space (i.e. sparse)
2. An array of hashrefs - this could contain something like an object. Each hashref could have the keys "x", "y" and "beans" and those values individually.
3. Some combination of the two, e.g. an array of hasherefs of arrayrefs . Keys being "coordinates" (value of arrayref) and "beans" ( value being scalar)

How you want to get your data back is important in determining how you should store it. A simple array of arrays might suit you:

my @arr;
while (my ($x, $y, $beans) = read_grid_cell) {
    push @arr, [$x, $y, $beans];
}
# Or the equivalent map statement, if you like.
[download]

my %hash;
while (my ($x, $y, $beans) = read_grid_cell) {
    $hash{$x}{$y} = $beans;
}
[download]

my @beans;
my %grid;
while (my ($x, $y, $beans) = read_grid_cell) {
    push @beans, [$x, $y, $beans];
    $grid{$x}{$y} = $beans;
}
# Now you can walk through @beans, or you can lookup by $x and $y,
# but don't change the beans in either spot, because they're not linke
+d!
[download]

So the input is unordered, each line indicates one bean at the location given, and you just want to tally them up and then print them out in grid order? Sounds like:

my %beancounts;
while (<>) {
    my ($x, $y) = split;
    ++$beancounts{$x}{$y};
}
for my $x (sort {$a <=> $b} keys %beancounts) {
    for my $y (sort {$a <=> $b} keys %{$beancounts{$x}}) {
        print "$x $y $beancounts{$x}{$y}\n";
    }
}
[download]

---

Caution: Contents may have been coded under pressure.

If you need real coordinates, an Array wouldn't be enough. You can switch to a HoH (Hash of hashes), something in the lines of:

my $HoH = { 101.67 => { 80   => 5 } , 
            312    => { 67.8 => 7 } };
[download]

Then you can retrieve the values with $HoH->{312}{67.8}.

Just be aware that hash keys are strings, so use sprintf if you get them from variables, or you're going to get strange results.

the other approach would be to simply run through the file, one line at a time, and to split() the line into $x, $y and $beans, after which you simply add it to a 1-dimensional hash using ie $beancounter{$x}{$y} = $beans;

hope that helps a bit