Vectorize for-loop with indexing

% I have following variables
B = [
    0 -0.5000 -0.7222 -0.7778 -1.1111;
    -0.2292 0 0 0 -1.1111;
    -0.6111 -0.5000 -0.4815 -0.5185 0
    ];

idx = [
    1 1 2 1 1;
    2 2 3 4 4;
    3 4 5 5 5
    ];

% I created for-loop
for i = 1:5
    sum(B(idx == i))
end

% And output is
% ans =
% 
%    -2.3889
% ans =
% 
%    -0.9514
% ans =
% 
%    -0.6111
% ans =
% 
%    -1.6111
% ans =
% 
%     -1

% I want to vectorize this for-loop and get following array
% ans =
%    -2.3889
%    -0.9514
%    -0.6111
%    -1.6111
%    -1.0000

Answer 1

[dpb]

 S=accumarray(idx(:),B(:));

See accumarray for all the many permutations that can be made...

An alternative using grouping variables in Statistics TB could be

 S=grpstats(B(:),idx(:),@sum);

or there's always splitapply

[zeraye]

Works perfectly! Thanks

[dpb]

Of course it does... <VBG>