[R] counting strings of identical values in a matrix

[Marc Schwartz]

On Thu, 2007-11-15 at 17:53 +0100, A M Lavezzi wrote:
> thank you.
> I did not think about the case of overlapping of 
> 1's from the end of one column to the start of the next,
> this would actually be a problem
> 
> In the simulations I am running each column 
> corresponds to the path followed by an agent 
> across states of a stochastic process,
> so I would like to avoid mixing up two different 
> paths (I made a mistake when I mentioned the possibility of turning my matrix
> into a vector, sorry about that).
> 
> can I kindly ask again your help on this?
> 
> please excuse me.
> 
> Mario

<snip>

Not a problem.  After sending my follow up, I suspected that you might
need a more general approach. This sort of ends up being a combination
of the first two, in order to keep each column sequence intact:


res <- do.call(cbind, apply(prova, 2, 
                            function(x) do.call(rbind, rle(x))))

> res
        [3,] [5,] [7,]   [4,] [7,]   [4,] [8,]   [2,] [3,]  
lengths    2    2    2 2    3    3 2    3    4 1    1    1 6
values     3    1    3 1    3    1 3    3    1 3    3    1 3



> table(res["lengths", res["values", ] == 1])

1 2 3 4 
1 2 1 1 


I think that should do it, but you might want to test it on a known set
of data.

HTH,

Marc