Problem with while loop, index is set back

I try to compare two line vectors in a way that the following is supposed to happen:

As long as a "0" is detected in vector1, then the index in the result-vector u is also "0". As soon as a "1" is detected, the value of vector2 is to be written to "u" until another "1" is detected in vector1.

The following example should illustrate what I mean:

vector1 = [0 0 0 0 1 0 0 0 0 0 1 0 0 0 0];

vector2 = [0 0 1 1 1 1 1 1 0 1 1 1 1 0 0];

The result should be:

u = [0 0 0 0 1 1 1 1 0 1 1 0 0 0 0]

Instead I get:

u = [0 0 0 0 1 0 0 0 0 0 1 1 1 0 0]

My code:

u = zeros(1,size(vector1,2));

for i = 1:size(vector1,2)

    if vector1(1,i) == 0
        u(1,i) = 0;
    else
        u(1,i) = 1;
        i = i+1;
        while vector1(1,i) <1
            u(1,i) = vector2(1,i);
            i = i+1;
        end
        u(1,i) = 1;
    end
end

I ran the debugger and the following happens: At i = 5 the first "1" is detected and the else statement with while loop starts. As soon as it enters the while loop, i is increased correctly. When the second "1" is detected in vector1, it exists the while loop correctly. u looks like this:

u = [0 0 0 0 1 1 1 1 0 1 1 0 0 0 0]

i = 11

But then, on the next turn, starting with the next "if" statement at the beginning, i is set back to i = 6 as if the while loop was never entered. Then the already existing spots are overwritten by "0" again. Probably something simple with overwriting i , but I have been trying to figure it out for the past 2 hrs and just couldn't solve it.

Thank you for your help!

Answer 1

[Roger Stafford]

It is not clear whether at the second '1' in vector1 of your example the '1' that is put into 'u' is copied immediately from vector1 or there is a delay by one position and it is still being copied from vector2. I am assuming the former of these. I also assume that copying from the two vectors continues to alternate between them as 1's alternately occur in vector1.

Try this (call vector1 and vector2, v1 and v2, resp.):

   f = find(v1>0);
   f1 = f(1:2:end);
   f2 = f(2:2:end);
   u = zeros(size(v1));
   u(f1) = 1;
   u(f2) = -1;
   u = cumsum(u);
   u = v1.*(u==0)+v2.*(u==1);

[Benedikt]

Yes, exactly. Copying from these two vectors continues to alternate between them as the 1's alternately occur in vector1. Thank you for your help, it works fine with your code! I really appreciate it.

Answer 2

[Guillaume]

You can't overwrite i. for loops in matlab do not work like that. At each iteration of the for, i takes the next value in the vector [1 2 3 ... size(vector1, 2)].

If you do want to your computation with a for loop, I would just use a flag to tell me whether to copy or not the element at each iteration. You toggle the flag whenever you encounter a one in vector1:

u = zeros(size(vector1));
docopy = false;
for idx = 1 : numel(vector1)
   if vector1(idx) == 1
      u(idx) = vector2(idx); %always copy when 1
      docopy = ~docopy; %toggle flag
   else
      u(idx) = docopy & vector2(idx); %will always be 0 if docopy is false
   end
 end