Message-ID: <Pine.A41.4.33.0105040829370.73704-100000@homer25.u.washington.edu>
Date: 2001-05-04T15:40:48Z
From: Thomas Lumley
Subject: scoping error in xy.coords (PR#932)
In-Reply-To: <200105041522.RAA11759@pubhealth.ku.dk>
On Fri, 4 May 2001 kjetilh@umsanet.edu.bo wrote:
> Hola!
>
> > rm(list=ls(all=TRUE))
>
> > x <- 1:20
> > y <- 1+x+rnorm(x)
>
> > xy.coords(y ~ x,NULL)
>
> ... expected output, correct, but when called from inside lowess:
>
> > lowess(y ~ x)
> Error in xy.coords(x, y) : x and y lengths differ
If you don't call them x and y it works
> x<-1:10
> y<-1:10
> lowess(y~x)
Error in xy.coords(x, y) : x and y lengths differ
> z<-y
> w<-x
> lowess(z~w)
$x
[1] 1 2 3 4 5 6 7 8 9 10
$y
[1] 1 2 3 4 5 6 7 8 9 10
The problem is that the formula is evaluated in the parent frame, where y
evaluates to NULL, and x evaluates to y~x, rather than in the global
environment where they would both evaluate to vectors.
In order to handle formulas correctly I think xy.coords needs to use the
same sort of indirect evaluation as model.frame: eg
xy <-substitute(xy.coords(X,Y),list(X=substitute(x),Y=substitute(y)))
xy<-eval(xy,parent.frame())
works as a replacement for the first two lines of lowess(). I hadn't
realised that xy.coords() was supposed to handle formulas, and since it
doesn't have a data= argument it can't do it very well.
This seems a bit drastic, given the number of times xy.coords is used.
Perhaps there's a simpler solution.
-thomas
Thomas Lumley Asst. Professor, Biostatistics
tlumley@u.washington.edu University of Washington, Seattle
^^^^^^^^^^^^^^^^^^^^^^^^
NOTE NEW EMAIL ADDRESS
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel 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-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._