[R] Interaction versus combinations

[Marc Schwartz]

On Jun 3, 2010, at 2:08 PM, Alex van der Spek wrote:

> I can get the interactions between factors like this:
> 
> > idx=c(1,3,6,9)
> > jdx=idx
> > levels(interaction(idx,jdx,lex.order=TRUE))
> [1] "1.1" "1.3" "1.6" "1.9" "3.1" "3.3" "3.6" "3.9" "6.1" "6.3" "6.6" "6.9"
> [13] "9.1" "9.3" "9.6" "9.9"
> 
> This list contains all possible interactions. Whereas I need only the combinations, e.g 4 over 2 = 6 total being
> 
> "1.3 " "1.6" "1.9" "3.6" "3.9" "6.9"
> 
> I have been unable to find how to get the combinations. Of course with some string manipulations this can be done for this toy example. There should be something in R to make this easy?
> 
> Thanks in advance
> Alex van der Spek


> combn(c(1, 3, 6, 9), 2)
     [,1] [,2] [,3] [,4] [,5] [,6]
[1,]    1    1    1    3    3    6
[2,]    3    6    9    6    9    9


See ?combn

HTH,

Marc Schwartz