Number of times the rows of a small matrix appear amongst the rows of a larger matrix, vectorization

Hi!

I need to find the number of times the elements in the rows of a small (Sx2) matrix appear in a larger (Lx4) matrix. I would like to do this in a faster way than doing it with a loop if possible. Some properties the matrices have:

• both matrices have unique elements on their rows and the order of appeareance doesn't matter
• an element from the small matrix can always be found at least once in the large matrix

For example:

L = [1     2     3     4
     2     5     7     4
     2     6     8     3
     3     1     2     8
     8     6     4     2];
S = [2 1
     2 4
     5 3];

for kk = 1:length(S(:,1))
times_each_row_appears(kk,1) = nnz(sum(ismember(L,S(kk,:)),2)==2); % sum(_,2) acts along the rows, sum==2 when both elements are found
end
times_each_row_appears

returns

times_each_row_appears =

     2
     3
     0

Does anyone have any suggestion on how to do this faster without a loop, preferrably on matrices with thousands of rows? Thank you!

[madhan ravi]

You forgot to post the expected result.

[Adri]

For the matrices in the example, the result should be

result =

     2
     3
     0

Answer 1

[madhan ravi]

[m, n] = size(S);
times_each_row_appears = zeros(m,1); %preallocate for speed!
for kk = 1:m
    times_each_row_appears(kk,1) = nnz(sum(ismember(L,S(kk,:)),2)==n);
end

Alternative is to use arrayfun(...) but loop is the fastest anyhow!