[R] Factor levels.

[Rolf Turner]

On 3/10/2007, at 5:10 PM, Christos Hatzis wrote:

> Would
>
> levels(fff) <- c("A","S","U")
>
> not work?

	Well, not quite.  This would scramble the levels.  The levels of the  
original fff are
	c("Unit","Achieved","Scholarship")  --- i.e. they are ***not*** in  
alphabetical order.

	But your suggestion modified to

		levels(fff) <- c("U","A","S")

	does indeed seem to work.

	I had a clear and distinct recollection of having tried something  
like this previously
	and getting a factor consisting entirely of NAs (since none of the  
assigned levels
	matched any of the existing levels).  But my clear and distinct  
recollection appears
	to have been a load of dingos' kidneys.

	I now believe that what I was recollecting was the fact that

		fff <- factor(fff,levels=c("U","A","S"))

	gives a factor consisting of nothing but NAs.  Subtly (???) different.

	I have a nagging feeling that there is some risk associated with  
your suggestion,
	as compared with Peter Dalgaard's.  But I can't see what that risk  
would actually be.

				cheers,

					Rolf Turner


######################################################################
Attention:\ This e-mail message is privileged and confidenti...{{dropped}}