3D Array Rastering

I am attempting to raster scan DMD mirrors by inputing a 3D array. I would like the array (for the number of mirrors along one side, n = 2) to look like: [[1,0;0,0],[0,1;0,0],[0,0;1,0],[0,0;0,1]] (i.e. a row major order scan of the mirrors). I intend to scale this array and my initial attempt hasn't been succesful. How do I create this array?

N = 4;
Array3D = zeros(2,2,N);

for row=1:2
    
for col=1:2
    
for k=1:N
    
  Array3D(row,col,k) = 1;
  
end
end 
end

[darova]

What about eye?

Answer 1

[Matt J]

I'm not sure from your description how this should generalize for different N, but perhaps this is what you want:

>> N=2; Array3D=reshape(eye(N^2),N,N,[])

Array3D(:,:,1) =

     1     0
     0     0


Array3D(:,:,2) =

     0     0
     1     0


Array3D(:,:,3) =

     0     1
     0     0


Array3D(:,:,4) =

     0     0
     0     1

[Matt J]

Incidentally, for large N, the array you are trying to building (assuming I've understand the goal correctly) will be highly RAM-consuming. You can reduce memory consumption by using my ndSparse class (Download). E.g.,

>> N=30; Array3D=reshape(eye(N^2),[N,N,N^2]); SparseArray3D=ndSparse(speye(N^2),[N,N,N^2]);

>> whos *Array3D
  Name            Size          Kilobytes     Class      Attributes
                                                                   
  Array3D         30x30x900          6329     double               
  SparseArray3D   30x30x900            22     ndSparse             
                                                                   
>> isequal(Array3D,SparseArray3D)

ans =

  logical

   1

[Dillon Kopecky]

Excellent! Thank you! I knew I must have been overcomplicating the solution. Thanks for the heads up regarding optimization and providing your code.

I needed row major operation which may not have initially been clear. The following code yields the correct ordering in case someone is interested.

N=3; 
A = eye(N^2);
Array3D=reshape(A',N,N,[]);
B = permute(Array3D,[2 1 3]);