Replace loop to create a matrix

Hello,

I want to create a matrix using the following code:

groupid = [1; 1; 1; 2; 2; 3; 3;]; 
data.id = groupid;
denominador(data)

ans = 
   (1,1)        1
   (2,1)        1
   (3,1)        1
   (2,2)        1
   (3,2)        1
   (3,3)        1
   (4,4)        1
   (5,4)        1
   (5,5)        1
   (6,6)        1
   (7,6)        1
   (7,7)        1

function f = denominador(X)
  [GC, GR] = groupcounts(X.id);
  args = cell(size(GR, 1));
  for i = 1:size(GR, 1)
    args{i} = sparse(tril(ones(GC(i))));
  end

  f = blkdiag(args{:});
end

The matrix created is C:

As you can see above, I am creating a sparse matrix where the diag is composed of lower triangular arrays of ones with dimensions equal to the number of times each "id" appears, and the elements outside the diag are zero.

Is there any function that can replicate in a more efficient way (without a loop) the function "denominador"? Is there any function similar to "sparse" than can be used to create matrix C?

Answer 1

[Matt J]

This might be a more efficient way to build arg, but I suspect that your bottleneck will be in blkdiag().

function f = denominador(X)

    GC = groupcounts(X.id);

    args=arrayfun(@(i) sparse(tril(ones(i))) ,1:max(GC),'uni',0 );

    args=args(GC);

    f = blkdiag(args{:});

end

[Adrián Martínez Rodríguez]

Thanks! It works.