my %hash;
my @split = split //, $string;
my @positions = map { [$split[$_], $_] } 1 .. @split-1;
undef @split;

foreach (@positions){
  push @{$hash{$_->[0]}}, $_->[1];
}
undef @positions ;
[download]

now, for each character you have a hash entry with all its positions in the string (array).
if for example you want to get the 3rd occurance of 'A', you will simply need to access: $hash{A}[2]
(2 will be the 3rd index in the array)

You wannted efficient function, and this way allows you to get you data in one hash+array access which IMHO is efficient.
the price you will pay is the one-time building of the hash (per string).

Enjoy,
Mickey