pulling duplicate data from a complex structure

[tcf03]

I have the following code

push ( @{$map{$$row[1]}{$$row[2]}}, $$row[0] ) unless $map{$$row[1]}{$
+$row[2]}{$$row[0]}++;

if I pull ou the unless statement -it all works fine - I just end up with duplicate data in my array. Can anyone suggest how to skip the duplicates? Ive tried several different bits of code:

push ( @{$map{$$row[1]}{$$row[2]}}, $$row[0] ) unless ${$map{$$row[1]}
+{$$row[2]}{$$row[0]}}++;

etc... and none seem to work.

thanks

UPDATE
I figured it out...

push ( @{$map{$$row[1]}{$$row[2]}}, $$row[0] ) unless grep { /$$row[0]
+/ } @{$map{$$row[1]}{$$row[2]}};

[injunjoel]

Greetings,
Though I would suggest strongly that you consider using a hash, here is something you can do after your pushes to de-dup your data.
Leave out the unless, since that works fine, and run the following after.

my @non_duplicates = do{
    #lets use a hash since we want unique values
    #and localize it so its a fresh copy
    local %_;

    #undef the values of this localized hash
    #while assigning your values as keys
    undef @_{@{$map{$$row[1]}{$$row[2]}}};

    #call keys on the hash to get a list of your
    #unique values
    keys %_;
}
#assign our unique list back to your original structure.
$map{$$row[1]}{$$row[2]} = \@non_duplicates;

Hope that helps

-InjunJoel

[Fletch]

Your update code scans the entire results so far for every new row added ("Can you say 'slow', boys and girls? Good, I knew you could</Fred>). Your first item doesn't work because you're trying to push onto something that isn't an array reference (thanks to your incrementing it). You want a parallel %seen hash which tracks what items you've already pushed.

[planetscape]

You may also wish to take a look at these nodes:


How to find and remove duplicate elements from an array?
How can I display a unique array if it contains some repeated elements

HTH,

[NiJo]

Having to remove duplicates manually raises a red flag. In most cases this means that you should redesign the data structure to use a hash.

The classic duplicate elimination converts the array to hash to array.