Indeed. If one desires for [1,2,3] and [3,2,1] to be considered identical, they must have identical signatures, and that can be achieved by sorting the numbers.

I doubt that's going to be a common desire, but if it happens, it's probably cleaner to normalise the data before finding the dups.

# Normalise the data.
@$_ = sort { $a <=> $b } @$_ for @AoD;

# Filter out duplicates.
my %seen;
@AoD = grep !$seen{ join(',', @$_) }++, @AoD;
[download]