[R] integer codes of factors
Spencer Graves
spencer.graves at pdf.com
Fri Jul 15 05:26:38 CEST 2005
What's the problem? As suggested by the help page, the numeric codes
are assigned in the order the names appear in the levels argument.
Consider the following example from the help page plus a minor
modification:
> (ff <- factor(substring("statistics", 1:10, 1:10), levels=letters))
[1] s t a t i s t i c s
Levels: a b c d e f g h i j k l m n o p q r s t u v w x y z
>
> as.numeric(ff)
[1] 19 20 1 20 9 19 20 9 3 19
> (ff <- factor(substring("statistics", 1:10, 1:10), levels=letters[3:1]))
[1] <NA> <NA> a <NA> <NA> <NA> <NA> <NA> c <NA>
Levels: c b a
> as.numeric(ff)
[1] NA NA 3 NA NA NA NA NA 1 NA
I highly recommend Venable and Ripley (2002) Modern Applied
Statistics with S (Springer) and V & R (2000) S Programming (Springer).
spencer graves
Mike R wrote:
> U = c("b", "b", "b", "c", "d", "e", "e")
>
> F1 = factor( U, levels=c("a", "b", "c", "d", "e") )
>
> as.numeric(F1)
> [1] 2 2 2 3 4 5 5
>
> Here, the integer code of "b" in F1 is 2
>
> K = factor( levels(F1) )
> as.numeric(K)
> [1] 1 2 3 4 5
> K
> [1] a b c d e
> Levels: a b c d e
>
> And again, the integer code of "b" in K is 2. Great!
>
> I am wondering how modify that usage such that the correspondence between
> the two numeric vectors can this be trusted. for example, the correspondence
> can be corrupted by placing the "a" at the end:
>
> F2 = factor( U, levels=c("b", "c", "d", "e", "a") )
>
> as.numeric(F2)
> [1] 1 1 1 2 3 4 4
>
> Placing the "a" at the end changed the integer code of "b" in F2 to 1, which is
> not a problem. But ......
>
> K = factor( levels(F2) )
> as.numeric( K )
> [1] 2 3 4 5 1
> K
> [1] b c d e a
> Levels: a b c d e
>
> But the integer code of "b" in K is now 2, which does not correspond to its code
> in F2.
>
> One would think that ordered=TRUE ought to avoid the corruption, but it does not
> seem to accomplish that:
>
> K = factor( levels(F2), ordered=TRUE )
> as.numeric(K)
> [1] 2 3 4 5 1
> K
> [1] b c d e a
> Levels: a < b < c < d < e
>
> But the integer code of "b" in K is still 2.
>
> However, corruption can be avoided with this idiom:
>
> K = factor( levels(F2), levels=levels(F2) )
> as.numeric(K)
> [1] 1 2 3 4 5
> K
> [1] "b" "c" "d" "e" "a"
> Levels: b c d e a
>
> Now the integer code of "b" in K is 1, which, as desired, is in
> correspondence with
> its code in F2.
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
--
Spencer Graves, PhD
Senior Development Engineer
PDF Solutions, Inc.
333 West San Carlos Street Suite 700
San Jose, CA 95110, USA
spencer.graves at pdf.com
www.pdf.com <http://www.pdf.com>
Tel: 408-938-4420
Fax: 408-280-7915
More information about the R-help
mailing list