Summing common elements of a matrix together

Suppose I have a matrix

A = [1 1 2 2 3 3 3]

and

B = [1 2 3 4 5 6 7]

I want to create a matrix which sums the elements of B, by grouping them according to A, which means:

sum = [3 7 18]

which we got by: from B, we took 1, 2 because A matrix tells us that first two elements belong to first group, similarly 3, 4 for second and 5, 6, 7 for the third. Is there a non-loop way to do this?

Answer 1

[Oleg Komarov]

accumarray(A.',B.')

[Geoff]

I've always used the syntax A' to transpose, but just checked the help for transpose() and see I'm apparently wrong! Is A' just an alias for A.', or is there some obscure difference?

[Oleg Komarov]

You're not wrong, it just matters with complex numbers:

' -> ctranspose: complex conjugate transpose

.' -> transpose

Answer 2

[Onomitra Ghosh]

May be something like:

A = [1 1 2 2 3 3 3];
B = [1 2 3 4 5 6 7];
C = [];
for idx = unique(A)
    C(end+1) = sum(B(A==idx));
end
C

[Vinay Pandey]

Is there a way to remove the loop?

[Onomitra Ghosh]

I have not tried it; it might be difficult to get rid of the loop because it is not straightforward vectorization.