Can anyone simplify this code please?

Morning all,

I have a problem. I'm trying to take a subset of a matrix, but not a uniform distribution. The difference between points is given in the vector phi, which I've made super long to make sure I don't reach the end of it, As you can see this vector is a 6 number repeating pattern. I've made a vector, zeta, containing the rows I would like to extract from the matrix. But, it is super clunky! Is anyone able to simplify this code for me please? I'm sure it should be possible without the loop.

phi = repmat([8.70;8.70;8.70;8.70;8.70;16.59],10,1);
% Random number generator to decide which sensor to use first
sensor = randi(6,1);
col = 1;
zeta = 1;
while col + round(phi(sensor),0) < 153
  col = col + round(phi(sensor),0);
  zeta = [zeta; col];
  sensor = sensor + 1;
end

Many thanks

[jlt199]

I said rows, clearly I mean columns

[John D'Errico]

Why? Does it take too long to run? pre-optimizing code that has no reason to be optimized is a huge waste of time.

[jlt199]

More as a learning exercise. I use too many loops in Matlab and don't take advantage of the vectorization enough. I want to start getting better at it before I start on the project with huge amounts of data to process

[jlt199]

Also, my boss reads Matlab so would be good to have sophisticated looking code ;)

[jlt199]

Does no comments mean that the code is not easily simplifies and I'm ok using a loop in this context? Thanks

Answer 1

[TallBrian]

I am not sure exactly what you mean by simplify but here is some code which eliminates the looping.

LENGTH = 153;
phi = round([8.70; 8.70; 8.70; 8.70; 8.70; 16.59]);
max_reps = round(LENGTH / sum(phi)) + 1;
phi_complete = repmat(phi, max_reps, 1);
phi_offset = [1; phi_complete(randi(6,1):end)];
zeta = cumsum(phi_offset);
zeta = zeta(zeta < LENGTH);

[jlt199]

Thanks Brian, this helps a lot