R help - Jul 2005 - integer codes of factors

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.

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