Skip to content

extract element, row, etc from general array structure

3 messages · Jason Turner, Peter Dalgaard, Thomas Lumley

Original
Jason Turner
# 
This one has me a bit stumped, and I know there may well be a really
simple way of doing it in R.  Sorry for the long-winded explanation.

Short version:
given an array of arbitrary number of dimensions D, how can we 
retrieve the Nth element of dimension D without knowing  in
advance what D will be?  For example, given that array
A has dimensions c(3,4,5), and N=2, can we write a function
that will "know" to return A[,,2], and not mistake that
for A[2] or A[,2]?  Bonus question - can such a function
be written without relying purely on an arbitrarily high
number of "if" statements?

Long version:
I've been using Greg Warnes' wapply function (thanks Greg!),
and modifying it to suit my own ends.  wapply() takes a vector,
and repeatedly applies a function FUNC (decided at run-time) to 
windows of the data.  The modification I've made is to generalize
the return structures - if a scalar is returned by FUNC, wapply
returns a vector.  If FUNC gives a vector, wappaly returns a
matrix.  etc...

A complementary function I'm trying to build is one that extracts
the results of FUNC for window N.  As I see it, the logic would 
be something like this:  

if:
foo<-wapply(some.time.series, some.vector.returning.function)

foo is a matrix.  The "extract" function would grab the column
of foo for window N.  e.g. N=5, extract would return foo[,5]
(maybe even with classes set appropriately, if I get particularly
clever).

Since we don't know in advance how many dimensions foo will have,
and the last dimension is the one we're interested in, how do
we express in R:
  if it's a vector, get foo[5],
  if it's a matrix, get foo[,5],
  if it's a 3d array, get foo[,,5], 
  ...
without being so ugly or having an arbitrary upper limit based on how
many "if" statements I was prepared to type?

If there's no elegant way, I'll work around it by keeping records
of data size, and pushing structures back to vectors when I 
need to automatically pull segments of data out.  If there's an 
easier way, I'd be most appreciative.

Cheers

Jason
Peter Dalgaard
# 
Jason Turner <jasont at indigoindustrial.co.nz> writes:
Something like this 

apply(foo, seq(length=length(dim(foo))-1),"[",5)

Won't work for vectors though since they don't have dim() and 1-D
arrays are probably also troublesome.
1 day later
Thomas Lumley
#
On Mon, 29 Oct 2001, Jason Turner wrote:

            

      
      
      
    

        
function(A,N){
 dims<-dim(A)
 D<-length(dims)
 skip<-prod(dims[seq(length=D-1)])
 slice<-A[ (N-1)*skip+1:skip]
 if (D>1) dim(slice)<-dims[seq(length=D-1)]
 return(slice)
}


	-thomas

Thomas Lumley			Asst. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._