[R] can I have a matrix of factors, please?

[Liaw, Andy]

> From: Liaw, Andy [mailto:andy_liaw at merck.com] 
> 
> Dear R-help,
> 
> Is it a violation of the S language to have a matrix of 
> factors?  What I would like to have is just a factor object 
> that has dim attribute, and can be printed (and subsetted) 
> like a matrix; i.e., all columns/rows have the identical 
> levels.  However, I can't get it to work:
> 
> > x <- factor(sample(2, 10, replace=TRUE))
> > dim(x)<- c(5,2)
> > x
>  [1] 1 2 2 1 2 1 1 1 2 1
> Levels: 1 2
> > str(x)
>  int [1:5, 1:2] 1 2 2 1 ...
>  - attr(*, "levels")= chr [1:2] "1" "2"
>  - attr(*, "class")= chr "factor"
> > x[1,]
> factor(0)
> Levels: 1 2
> 
> (This is R 1.7.1 on WinXPPro.)

This is wierder than I thought!  Try:

> x[,1]
[1] 1 2 2 1 2
Levels: 1 2
> x[2,]
factor(0)
Levels: 1 2

I.e., column extraction "works", but not row extraction.

> y <- factor(1:12)
> dim(y) <- c(3,4)
> y[,1]
[1] 1 2 3
Levels: 1 2 3
> y[1,1]
[1] 1
Levels: 1
> y[1,]
factor(0)
Levels: 1 2 3 4 5 6 7 8 9 10 11 12
> y[1, 1:ncol(y)]
[1] 1  4  7  10
Levels: 1 4 7 10
Warning message: 
the condition has length > 1 and only the first element will be used in: if
(drop) factor(y) else y 

Column extraction and element extraction "work", but notice the unused
levels mysteriously disappeared!  The last warning also seems strange.

> is.matrix(y)
[1] TRUE
> class(y)
[1] "factor"

This also baffles me:  Given a numeric vector, (say y <- 1:12, dim(y) <- 3:4
will give y the "matrix" class.  Yet if I do this with a factor, it doesn't!
Apparently that's not what is.matrix tests for.

> The alternative is to have a list instead, where each 
> "column" makes up a component of the list.  But it would be 
> nice to have the matrix...

Short of as.list(as.data.frame(matrix))), does anyone know of an efficient
way to turn a matrix into a list with each component containing a column?
 
Best,
Andy