R help - May 2004 - Lists and outer() like functionality?

Hi,

I'm have a list of integer vectors and I want to perform an outer() 
like operation on the list. As an example, take the following list:

mylist <- list(1:5,3:9,8:12)

A simple example of the kind of thing I want to do is to find the sum 
of the shared numbers between each vector to give a result like:

result <- array(c(15,12,0,12,42,17,0,17,50), dim=c(3,3))

Two for() loops is the easiest way but I wondered if there was a 
neater/faster solution.

mylist.len <- length(mylist)
ind <- 1:mylist.len
result <- array(NA, dim=c(mylist.len,mylist.len))
for(x in ind){
   for(y in ind){
     result[x,y] <- sum(mylist[[x]][test[[x]] %in% test[[y]]])
   }
}


Many thanks,
David Orme

David Orme <d.orme at imperial.ac.uk> writes:
> Hi,
> 
> I'm have a list of integer vectors and I want to perform an outer()
> like operation on the list. As an example, take the following list:
> 
> mylist <- list(1:5,3:9,8:12)
> 
> A simple example of the kind of thing I want to do is to find the sum
> of the shared numbers between each vector to give a result like:
> 
> result <- array(c(15,12,0,12,42,17,0,17,50), dim=c(3,3))
> 
> Two for() loops is the easiest way but I wondered if there was a
> neater/faster solution.
> 
> mylist.len <- length(mylist)
> ind <- 1:mylist.len
> result <- array(NA, dim=c(mylist.len,mylist.len))
> for(x in ind){
>    for(y in ind){
>      result[x,y] <- sum(mylist[[x]][test[[x]] %in% test[[y]]])
>    }
> }
> 
How about

> mysum <- function(x,y)mapply(function(x,y)sum(intersect(x,y)),x,y)
> outer(l,l,mysum)
     [,1] [,2] [,3]
[1,]   15   12    0
[2,]   12   42   17
[3,]    0   17   50


(notice that your problem really isn't that you have a list, but that
you have functions that don't vectorize over lists.)

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907

"David Orme" <d.orme at imperial.ac.uk> wrote:
...
> I'm have a list of integer vectors and I want to perform an outer()
> like operation on the list. As an example, take the following list:
>
> mylist <- list(1:5,3:9,8:12)
>
> A simple example of the kind of thing I want to do is to find the sum
> of the shared numbers between each vector to give a result like:
>
> result <- array(c(15,12,0,12,42,17,0,17,50), dim=c(3,3))
>
> Two for() loops is the easiest way but I wondered if there was a
> neater/faster solution.
>
> mylist.len <- length(mylist)
> ind <- 1:mylist.len
> result <- array(NA, dim=c(mylist.len,mylist.len))
> for(x in ind){
>    for(y in ind){
>      result[x,y] <- sum(mylist[[x]][test[[x]] %in% test[[y]]])
>    }
> }
>
I'm having a hard time figuring out what you want.  Who is this mystery
object "test"?  When I drop "test" from the text, and use:

result[x,y] <- sum(mylist[[x]] %in% mylist[[y]])

It doesn't give a syntax error, but the result is nothing like the one you
posted above.

As a style point, ind <- seq(along=mylist) is a bit more foolproof in the
case of empty lists being accidentally passed to the code (been bitten
that way once or twice).

Cheers

Jason