R devel - Sep 2000 - Re: more experience with formulas

>>>>> "WSt" == Werner Stahel
<stahel@stat.math.ethz.ch> writes:
[to me privately]

    WSt> I do not want to call this a bug, but it bugged me. Please try the
    WSt> following:

 WSt>  tform <- sqrt(RADAI) ~ sqrt(RADAI.e) + TAGE.ej + SPITAL + ARZT +
DAS28 +
 WSt>     SJC + TJC + DGA + HAQ + PGA + PAIN + YEAR.SYM + YEAR.DIA + 
 WSt>     ALTER + IndV + m.NSAIDs + m.Stereoids + DM.Cyclosp + DM.Penic
 WSt> 
 WSt>  as.character(tform) 

--> length-3 vector "~" "<LHS>"
"<RHS>".
However, the end of <RHS> is TRUNCATED!
and that truncation of the third part is a bug in my opinion.

The fact that it returns a vector of the same length as its argument is
very consistent behavior even though it maybe unexpected to the casual R user.

I think you'd want to use format() instead.
There, we can see that format.default really uses deparse(x) when x is such
a language object.
One could argue that the optional "width.cutoff = 60" argument to
deparse()
should also be settable from format(), in that case,
since
	format(tform) == deparse(tform)

really returns a vector of "3 lines" instead of just one line
which we'd get by the result of
     deparse(tfrom, width = 200)
       
Opinions?

 WSt> paste(tform,collapse="")# "y ~ x" becomes
"~yx" -- is S compatible
 WSt> as.character(tform[[3]])# again length-3 vector  "+"
"sqr..." "DM.Penic"
 WSt> 
 WSt> as.character.formula <- function(x, ...)
paste(deparse(x),collapse="")
 WSt> ## now ,
 WSt> as.character(tform)
 WSt> paste(tform,collapse="")
 WSt> ## work, but
 WSt> as.character(tform[[3]])
 WSt> ## still does not, since it is of mode language, but not of class
formula.
 WSt> ## It would be better to write as.character.default such that is works
for
 WSt> ## mode language.

It *does* work; it just returns a vector of character, again
fulfilling the semantic that
when
	    y <- as.character(x)
then
	  y[i] == as.character(x[i])

    WSt> PS: Thanks for your efforts in developping R.

    WSt> Werner Stahel

Martin Maechler <maechler@stat.math.ethz.ch>
http://stat.ethz.ch/~maechler/
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

On Tue, 19 Sep 2000, Martin Maechler wrote:
> >>>>> "WSt" == Werner Stahel
<stahel@stat.math.ethz.ch> writes:
> [to me privately]
> 
>     WSt> I do not want to call this a bug, but it bugged me. Please try
the
>     WSt> following:
> 
>  WSt>  tform <- sqrt(RADAI) ~ sqrt(RADAI.e) + TAGE.ej + SPITAL + ARZT
+ DAS28 +
>  WSt>     SJC + TJC + DGA + HAQ + PGA + PAIN + YEAR.SYM + YEAR.DIA + 
>  WSt>     ALTER + IndV + m.NSAIDs + m.Stereoids + DM.Cyclosp + DM.Penic
>  WSt> 
>  WSt>  as.character(tform) 
> 
> --> length-3 vector "~" "<LHS>"
"<RHS>".
> However, the end of <RHS> is TRUNCATED!
> and that truncation of the third part is a bug in my opinion.
I think using as.character() is the problem: this is what deparse() is
for.  Similar things happen in S too.  You may notice that in
versions of R up to and including 1.1.1 that anova methods had this bug:
I even filed a bug report against it.

I suspect it may be relatively unproblematic to change as.character because
where it matters the cognescenti (read, `the much-bitten') would
use deparse() anyway.

-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._