analog to the matlab buffer function?

library(gtools)
t(running(1:5, width=3, fun=c))

Here is another possibility, though I may be missing how the Matlab
function handles incomplete rows generated at the end of the source
vector. I have not fully tested this, so it may yet require some
tweaking and certainly appropriate error checking.

I am presuming that the basic premise is that each row is of length
'window' and that it overlaps with the END of prior row by 'overlap'.


Buffer <- function(x, window, overlap)
{
  Res <- NULL

  while (length(x) >= window)
  {
    Res <- c(Res, x[1:window])
    x <- x[(1 + window - overlap):length(x)]
  }

  matrix(Res, ncol = window, byrow = TRUE)
}

Buffer(1:5, 3, 2)

     [,1] [,2] [,3]
[1,]    1    2    3
[2,]    2    3    4
[3,]    3    4    5

Buffer(1:10, 4, 2)

     [,1] [,2] [,3] [,4]
[1,]    1    2    3    4
[2,]    3    4    5    6
[3,]    5    6    7    8
[4,]    7    8    9   10

HTH,

Marc Schwartz


On Thu, 2006-11-30 at 16:32 -0800, Charles C. Berry wrote:

See

?embed

It is not quite the same, but this seems to be what you want - at least
for the example you give:

t( embed(1:5,3) )[3:1,]

      [,1] [,2] [,3]
[1,]    1    2    3
[2,]    2    3    4
[3,]    3    4    5

On Fri, 1 Dec 2006, Martin Ivanov wrote:

Hello! I am new to R. I could not find a function analogous to matlab's
function buffer, which is used in signal processing. Is there such a
function in R? What I need to do is as follows. If I apply the function
to the vector c(1:5) for example with a window length 3 and overlapping
2, I need to get a matrix like this:
1 2 3
2 3 4
3 4 5
In matlab this is achieved with the function buffer. Is there ananalogous R
function?