How can I change the elements of a matrix in new one?

I want to change the elements of a matrix with a loop command. What I have done is, let’s assume that we have a matrix A=[-1 -5 -2 3;-4 -7 8 -3;-2 -3 -1 5;4 8 2 -3]

[m,n]=size(A);
B=[]
for i = 1:m
  I = (i-1)*2;
  for j = 1:n
      J = (j-1)*2 + 0.5;
      if(A(i,j)>0)
B=[B [J;m-I]]
         end
end
end

and the results that I get is

B =
  6.5000
  4.0000
B =
  4.5000
  2.0000
B =
  6.5000
       0
B =
  0.5000
 -2.0000
B =
  2.5000
 -2.0000
B =
  4.5000
 -2.0000

How can I take all the numbers of B in one Nx2 matrix ?

[Stephen23]

Doing this in a loop is very poor use of MATLAB. Learn how to write vectorized code and take advantage of the neater, faster solutions to problems like this.

Answer 1

[Azzi Abdelmalek]

A=rand(4)
[m,n]=size(A);
B=[]
for ii = 1:m
  I = (ii-1)*2;
  for jj=1:n
      J = (jj-1)*2 + 0.5;
      if(A(ii,jj)>0)
          B=[B;J m-I]
      end
  end
end

[Stephen23]

When you write this kind of code in MATLAB it will come up with this warning message (which I clicked on to expand the details):

This is because expanding arrays is a really bad idea, and doing this is one of the main ways the beginners slow down their code. See my answer to know how much faster and neater MATLAB code can be, using best practices that do not result in any warnings about writing slow code. Why bother learning bad programming habits that you just have to change later when you can learn how to program properly from the start?

[Ang Vas]

Thank you a lot for your help and I am sorry for the mess!!!

Answer 2

[Stephen23]

Instead of slow and painful loops as if MATLAB was some low-level programming language like C, just use simple vectorized code, which is faster and neater:

A = [-1,-5,-2,3; -4,-7,8,-3; -2,-3,-1,5;4,8,2,-3];
[x,y] = find(A.'>0);
X = (x-1)*2 + 0.5;
Y = size(A,1) - (y-1)*2;
B = [X,Y].'

Which displays this in the command window:

B =
  6.5000    4.5000    6.5000    0.5000    2.5000    4.5000
  4.0000    2.0000         0   -2.0000   -2.0000   -2.0000