Same procedure for different matrices

Hey there, I'm not good with writing loops and I guess I need one for my problem. I have a couple of matrices of different sizes that I would like to split into quadrants and then take the average of all non-zero elements for each quadrant which I am able to do with the following code:

    B = A(1:(mod(end, 2)==1)*0.5+end/2, 1:(mod(end, 2)==1)*0.5+end/2);
    C = A(1:(mod(end, 2)==1)*0.5+end/2, (mod(end, 2)==1)*0.5+end/2+ 1:end);
    D = A(end/2+(mod(end, 2)==1)*0.5 + 1:end, 1:(mod(end, 2)==1)*0.5+end/2);
    E = A(end/2+(mod(end, 2)==1)*0.5 + 1:end, end/2+(mod(end, 2)==1)*0.5+ 1:end);

    avg_A(:,:) = [mean(B(B(:)>0)) mean(C(C(:)>0)); mean(D(D(:)>0)) mean(E(E(:)>0))];

where A is the first of many matrices. Now, my question is, whether I can apply that code to all my other matrices (A2,A3,A4,...) without copying this code and change all the "A"s to "A2"s and so on. In the end I would like to have a matrix of size 2x2xnumber of matrices.

Hope you have enough information and can help me with that.

[Stephen23]

Do NOT waste your time writing bad code that creates or acceses separate variables with numbered variable names: "A frequent use of the eval function is to create sets of variables such as A1, A2, ..., An, but this approach does not use the array processing power of MATLAB and is not recommended."

source: http://www.mathworks.com/help/matlab/matlab_prog/string-evaluation.html

Just use indexing: simple, fast, efficient, neat, simple, easy to understand, much less buggy, easier to debug, etc, etc, etc.

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

https://www.mathworks.com/matlabcentral/answers/143-how-do-i-make-a-series-of-variables-a1-a2-a3-a10

https://www.mathworks.com/matlabcentral/answers/57445-faq-how-can-i-create-variables-a1-a2-a10-in-a-loop

etc, etc, etc.

"whether I can apply that code to all my other matrices (A2,A3,A4,...) without copying this code"

Yes, when you put all of those arrays into one ND-array or cell array, and then simply loop over it using indexing. Note that your code may also be best put into a function.

[Felix Koenig]

First of all, thank you for your answer. Don't worry, I'm using indices, I just wrote A2,A3,.. for simplicity here.

I really am a matlab greenhorn, but as far as I know, I don't think I can put all my arrays into one, since they all have different dimensions. So A_1 would be a 23x24 matrix, A_2 would be 22x25 and so on. If there is a way to put them all into one array, please show me how to do that.

My plan was to split those matrices into quadrants first (if the number of rows or columns is odd, it's alright for one quadrant to be larger than another one) and then calculate an average for each quadrant. That way I get multiple 2x2 matrices that I can all put into one array.

Please correct me if that is wrong thinking.

[Jan]

Use a cell array instead as a container for arrays of different size:

A = {rand(23, 25), rand(22,25)};
size(A{1})
size(A{2})

Answer 1

[Jan]

Use a subfunction:

A = {rand(23, 25), rand(22,25)};
B = zeros([2, 2, numel(A)]);
for iA = 1:numel(A)
  B(:, :, iA) = QuadrantMean(A{iA});
end

function M = QuadrantMean(A)
S   = floor(size(A) / 2);
A11 = A(1:S(1),     1:S(2));
A12 = A(1:S(1),     S(2)+1:end);
A21 = A(S(1)+1:end, 1:S(2));
A22 = A(S(1)+1:end, S(2)+1:end);
M   = [mean(A11(A11 ~= 0)), mean(A12(A12 ~= 0)); ...
       mean(A21(A21 ~= 0)), mean(A22(A22 ~= 0))];
end