Need help inproveing function...

Hello Everyone, I have a function that looks at a vector of numbers and will mark any numbers in that vector that repeat with the value they repeat with. I have code that is working, but it is the ugly and really bad programing and I know it. If anyone could help out I would much appreciate it! Here is an example and the code that is bad that I'm using:

    x = [1 1 3 5 4 6 6 1 9 1 1 8 2 5 5 5 2 7 7 2 2 2];
    out = [2 0 0 0 2 0 0 2 0 0 3 0 2 3]
  %So in x (from the beginning): 
  %there are two 1's that become 2; 
  %then 3 mismatching values of 3,5, and 4 that become 0, 0, and 0;
  %then two 6's that become 2;
  %then 2 mismatching values of 1 and 9 that become 0 and 0;
  %then two 1's that become 2;
  %then 2 mismatching values of 8 and 2 that become 0 and 0;
  %then three 5's that become 3;
  %then 1 value of 2 (no matching neighbors) that become 0;
  %then two 7's that become 2;
  %then three 2's that become 3;

%code that works this way:

      x = [1 1 3 5 4 6 6 1 9 1 1 8 2 5 5 5 2 7 7 2 2 2];

      x = [x 0];
      c = 1;
      num = x(1);
      b = 2;
      r = 1;
      for n = 2:length(x)
          if b >= length(x)
              break 
          end
          if x(b) == num
              if n == length(x)
                 break 
              end
              while x(b) == num
                  if b >= length(x)
                      r = r + 1;
                      g(c) = r;
                      break 
                  end
                  b = b + 1;
                  num = x(b - 1);
                  r = r + 1;
              end
              g(c) = r;
              c = c + 1;
              r = 1;
          else

              while x(b) ~= num
                  b = b + 1;
                  num = x(b - 1);        
                  if b >= length(x)
                      break 
                  end
                  if x(b) ~= num
                      g(c) = 0;
                      c = c + 1;
                  end  
              end
          end
      end
    out = g
      %output:
      out = [2 0 0 0 2 0 0 2 0 0 3 0 2 3]

Well that is the really ugly function, any help with it will be great!
Thanks!

Answer 1

[Evan]

Does this do what you want?

>> x = [1 1 3 5 4 6 6 1 9 1 1 8 2 5 5 5 2 7 7 2 2 2];
>> idx = cumsum([1 diff(x) ~= 0]);
>> out = accumarray(idx.',ones(size(idx))).';
>> out(out == 1) = 0;
out =
     2   0   0   0   2   0   0   2   0   0   3   0   2   3

For run-length encoding (i.e. what it looks like you're doing here) the cumsum and accumarray functions are very helpful, although they can be tricky to get used to at first.

[Chris E.]

Yes! Thank you a bunch! That is so much better. Thank you!

[Evan]

No problem! =)