How do I decrease the time taken by these two for loops. I am taking 30 seconds approximately to run this, and this is in a big for loop that runs for 10000 times.

Could anyone please help me with this? I am new to using MATLAB so I have written the code in a layman fashion and I am sure that there are ways to write it better.

The code is below (takes about 30 seconds, depending on size of both the matrices)

for i = 1:1:size(binary_probability_matrix,1)
    for j = 1:1:size(matrix,1)
        if matrix(j,1:2) == binary_probability_matrix(i,1:2)
            matrix(j,3:4) = binary_probability_matrix(i,3:4);
        elseif matrix(j,1) == binary_probability_matrix(i,2) && matrix(j,2) == binary_probability_matrix(i,1)
            matrix(j,3:4) = binary_probability_matrix(i,3:4);
        end
    end
end

Basically I am comparing the first two columns of matrix and binary_probability_matrix. If the 1st 2 columns in the matrix are 4 3 ..then I search in the first two columns of binary_probability_matrix for 4 3 OR 3 4.

Similarly, when I am searching for 10 18, I search the first two columns of the binary_probability_matrix for finding 10 18 OR 18 10.

I hope that I was clear in mentioning my question. Please help me with this as the execution time is quite large.

EDIT: order of matrix is 5000x4 and the order of binary_probability_matrix is 2500x7,

Thank you

Jay

[Turlough Hughes]

Do you have code to generate the matrices or can you upload the variables as a .mat.

[Jay Vaidya]

Yes I have a big code which generates these 2 matrices.

Answer 1

[Jay Vaidya]

Thank you Stephen Cobeldick. I used ismember() function and that is excuting this in no time.

binary_probability_matrix(:,1:2) = sort(binary_probability_matrix(:,1:2),2);

A1 = matrix;

[X,Y] = ismember(matrix(:,[1,2]),binary_probability_matrix(:,[1,2]),'rows');

matrix(X,3:4) = binary_probability_matrix(Y(X),3:4 );

isequal(A1,matrix); %just to check if they are really matched.