Relating values in one array indexed to another

I was given some arrays that have been indexed in a funny way. I think the idea was to save space because these are multi-million-point arrays.

Here's the compression and indexing method: I have a compressed array of values a_val and corresponding index record numbers a_ind. I also have a much longer array of indices b_ind. For every record index in b_ind there is a matching a_ind. I want to find the values b_val which will correspond to the values in a_val, but will be the same length as b_ind. Here's what I have:

a_ind = [ 2  3  5  6  7  9 11 12 13 15]; 
a_val = [10 10 20 30 40 40 40 50 60 60]; 
b_ind = [2 2 2 3 5 5 6 6 7 9 9 9 11 12 12 13 13 15];

I want:

b_val = [10 10 10 10 20 20 30 30 40 40 40 40 40 50 50 60 60 60];

I can do this with a loop:

b_val = NaN(size(b_ind)); 
for k = 1:length(b_val)
    b_val(k) = a_val(a_ind==b_ind(k)); 
end

but my arrays contain millions of elements and the loop takes several minutes to complete. I'm sure there is some clever way to use ismember, but I can't think of how. Any ideas?

[Chad Greene]

This works:

[~,lib] = ismember(b_ind,a_ind)
b_val = a_val(lib);

However, my b_ind dataset has a few indices not found in a_ind. If

b_ind = [2 2 2 3 5 5 6 6 7 9 9 9 11 12 12 13 13 15];

the ismember method throws a subscript index error because some lib values are zero.

Answer 1

[Chad Greene]

I figured it out.

[~,lib] = ismember(b_ind,a_ind)
b_val = NaN(size(b_ind));
b_val(lib>0) = a_val(lib(lib>0));