Efficient matrix comparisons that retain row/column order

Hi Everyone,

I've been struggling to figure out an efficient way to compare matrices and have the results stored in a matrix of a comparable size and similar ordering rather than in a vector of scalars for the case where the comparison is true. Can anyone share with me a faster way of doing this than just using for loops?

For example, suppose we start with:

mat1 = rand(10,10);
mat2 = rand(10,10);
difr = mat1 - mat2;

What I want to do is create a new matrix called mat3 (that is of the same dimensions as mat1 and mat2) where each element mat3(i,j) = mat1(i,j) if difr < 0.2 but 0 otherwise.

To do this, I make my best guess and try using the following code (whch I know is incorrect):

mat3 = zeros(10,10);
mat3 = mat1(diff < 0.2)

This turns mat3 into an Nx1 matrix where N is the number of times that the comparison is true ... and all of the information about position (e.g. i,j) in the original mat1 matrix is lost.

Can anyone share the proper way of doing this?

Thank you,

R

Answer 1

[per isakson]

This would be my first trial:

    mat3 = mat1;
    mat3( diff >= 0.2 ) = 0;

This isn't magic it is logical indexing!

[Rachel]

Thank you!