R help - Jun 1999 - coercing factors to matrix() --> num/char ? --

Peter Dalgaard BSA <p.dalgaard at biostat.ku.dk> wrote:
> Martin Maechler <maechler at stat.math.ethz.ch> writes:
[ snip... ]
> > >From this logic, coercing to a matrix should coerce to character
rather
> > than integer/numeric codes.
> 
> Hm. I'm not sure there really is much logic in this... Factors are
> sometimes character-like, sometimes integer-like.
I'd be interested to hear a discussion of when they need to
behave like integers.  In my work I can't ever remember needing
anything other than the character representation in my "user
level" coding.  I can see the need to access the integers for
"lower level" programming, for example, but when else?

(My thought has always been that factors should *always* coerce
to character unless the user explicitly requests otherwise.)

[ snip... ]
> Not really sure whether the problem is that R is consistent with
> itself or inconsistent with S. If we change as you suggest, we'll
> have two incompatibilities instead of one...
Factors are one area where I deem inconsistency with S a distinct
advantage.  :-)

--Todd
-- 
Z. Todd Taylor
Pacific Northwest National Laboratory
Todd.Taylor at pnl.gov
Why do you say there's one left, when it didn't leave?
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

"Z. Todd Taylor" <Todd.Taylor at pnl.gov> writes:
> I'd be interested to hear a discussion of when they need to
> behave like integers.  In my work I can't ever remember needing
> anything other than the character representation in my "user
> level" coding.  I can see the need to access the integers for
> "lower level" programming, for example, but when else?
> 
> (My thought has always been that factors should *always* coerce
> to character unless the user explicitly requests otherwise.)
The main one is probably indexing:
> ff<-factor(rbinom(20,1,.5),levels=0:1)
> x<-rnorm(20)
> y<-rnorm(20)
> plot(x,y,pch=c("f","m")[ff])
.which BTW also applies to the cbind issue since you can do things like:
> ff<-factor(rbinom(10,1,.5),levels=0:1)
> gg<-factor(rbinom(10,1,.5),levels=0:1)
> data.frame(ff,gg,comb.code=matrix(1:4,2)[cbind(ff,gg)])
   ff gg comb.code
1   0  1         3
2   1  0         2
3   1  0         2
4   0  1         3
5   0  1         3
6   0  0         1
7   1  1         4
8   0  0         1
9   1  0         2
10  1  1         4

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Peter Dalgaard:
>> Hm. I'm not sure there really is much logic in this... Factors are
>> sometimes character-like, sometimes integer-like.
Z. Todd Taylor:
>I'd be interested to hear a discussion of when they need to
>behave like integers.
Yesterday, using S+ 4.5 (Win), for instance:

#Exp is factor containing experiment numbers
>levels(mydata$Exp)
 [1] "0"  "1"  "2"  "3"  "4" 
"5"  "6"  "7"  "8"  "9" 
"10" "11" "12"
 [14] "13" "14" "15" "16" "17"
"18" "19"
>lm(y ~ x1 * x2, data=mydata, subset = Exp == 3 | Exp == 5,
+  weights = 5.5 - Exp)
Error in Ops.factor(5.5, Exp): "-" not meaningful for factors
Dumped
>lm(y ~ x1 * x2, data=mydata, subset = Exp == 3 | Exp == 5,
+  weights = 5.5 - as.integer(Exp) )
Error in lm.wfit(X, Y, weights, method): negative weights not allowed
Dumped
# . . . because as.integer(Exp) gives the codes as integers, not the
values, and
# the code for Exp 5 is 6.
> lm(y ~ x1 * x2, data=mydata, subset = Exp == 3 | Exp == 5,
+   weights = 5.5 - as.integer(as.character(Exp)))

Call: 
lm(formula = y ~ x1 * x2, data = mydata,
#. . . Remainder of regression report truncated.

Z.T.T.:
>I can see the need to access the integers for
>"lower level" programming, for example, but when else?
ALL of my programming is "lower level"!  ;-)
************************************************************
John J. Thaden, Ph.D., Instructor        jjthaden at life.uams.edu
Department of Geriatrics                     (501) 257-5583
University of Arkansas for Medical Sciences  FAX: (501) 257-4822
      mail & ship to:	J. L. McClellan V.A. Medical Center
		Research-151 (Room GB103 or GC124)
		4300 West 7th Street
		Little Rock AR 72205 USA
***********************************************************
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._