Finding the numbers of a vector with the same length?

Dear all

Assuming there is a vector consisting only 0 and 1 like:

x=[0 1 1 0 1 0 0 1 1 1 0 1 0 1 1 1 0 1 1 0 1 1 0 1]

I want to figure out how many times 1 or 1 1 or 1 1 1 or even longer 1 in a row exists are in this vector?

The solution here would be:

 nr_1= 3
 nr_1_1= 3
 nr_1_1_1= 2
 nr_1_1_1_1 = 0
 ....

Is there a easy way to do it?

Thanks for your help

(Matlab 2010a)

Answer 1

[C.J. Harris]

rep=diff(find(diff([-Inf x Inf])));
val=x(cumsum(rep));
out = hist(rep(val == 1), max(rep));

[Boris]

The main disadvantages about this solution is, that you will get a lot of zeros in the out-vector. Otherwise, it works well. Thanks a lot

[C.J. Harris]

You can process the out data if you want to remove the zeros, for example:

[out(find(out ~= 0)); find(out ~= 0)]

This will then show occurrences (top row) and sequence length (bottom row).

Answer 2

[Matt Fig]

Here's another solution:

D = diff(find([1 ~x 1]))-1;
D = histc(D,1:max(D))

[Andrei Bobrov]

+1

[Boris]

Works well. Thanks a lot.

Answer 3

[Harshit]

Hi,

You can save zero as it is and whenever 1 appears start a counter and count the number of 1s in succession. It is Run length encoding. When done just count the number of times the numbers appears.