unique elements in a Hash of Arrays?

[Anonymous Monk]

Dear Monks,
can you please guide me on how to find the number of unique elements in a hash of arrays?
I know that, if you have a normal array, you can just do:

$unique_ids = scalar(grep {defined $_} @normal_array);

but how can I do the same thing if there is a has of arrays involved? For example, how do you find the number of unique elements in the array for each of the keys in this example

     %HoA = (
    flintstones => [ "fred", "barney","barney", "nick", "john", "barne
+y" ],
    jetsons => [ "george", "jane", "elroy", "mike", "elias" ],
    simpsons => [ "homer", "marge", "bart", "jack", "homer" ],
    );

[AppleFritter]

Actually, this:

$unique_ids = scalar(grep {defined $_} @normal_array);

will not find the number of unique IDs in @normal_array, it'll simply count how many elements of @normal_array are defined. I recommend using the uniq function from List::MoreUtils.

Applying it to a hash of arrays is straightforward: just iterate through the keys, and apply it to each element of your hash (i.e., each array). Using your example:

foreach (sort keys %HoA) {
    say "$_: ", join ",", uniq @{ $HoA{$_} };
}

Output:

flintstones: fred,barney,nick,john
jetsons: george,jane,elroy,mike,elias
simpsons: homer,marge,bart,jack

If you can't or don't want to use this module, the Perl Cookbook has several solutions (Recipe 4.6, p. 102).

[AnomalousMonk]

I recommend using the uniq function from List::MoreUtils. ... If you can't or don't want to use this module ...

Anonymonk: ... just take a look at the code in it to see how it works; i.e., Steal This Code!

[Anonymous Monk]

Great thanks,
can you also tell me how I can count the unique elements per key of the Hash of Arrays?

[AppleFritter]

Sure. uniq returns a list of unique elements, and lists evaluate to their number of elements in scalar context, so if you replace the line inside the foreach with e.g. this:

    say "$_: ", scalar uniq @{ $HoA{$_} };

Then you'll get the number of unique items. (You could equally well assign to a variable for further processing, of course, and if you assign to a scalar, the explicit scalar won't be necessary, though it also won't hurt and may serve to make it a bit clearer what's going on.)