[Rd] Re: [R] NaN/NA and sprintf double format (PR#1561)

[J.C.Rougier@durham.ac.uk]

ripley@stats.ox.ac.uk wrote:
> 
> On Tue, 14 May 2002, Dirk Eddelbuettel wrote:
> 
> > With a (late prerelease of) 1.5.0 on Windows, using sprintf on NaN
> > yields garbage when the number of decimal is specified while it
> > works fine for NA:
> >
> > > sprintf("%.2f", NA)
> > [1] "0.00"
> 
> I get that on Linux too, but why is it fine?  See below.
> 
> > > sprintf("%.2f", NaN)
> > [1] "-1.#J"
> >
> > Not specifying the decimal precision lets both print garbage:
> >
> > > sprintf("%f", NaN)
> > [1] "-1.#IND00"
> > > sprintf("%f", NA)
> > [1] "-1.#QNAN0"
> >
> > Is this a bug or a feature?
> 
> Remember that NA is logical, so this is not correct usage.  Not that the
> correct usage works very much better ....
> 
> The C code says
> 
> /* Simple wrapper for C sprintf function: does very little checking to
>    see that the format conversion character and the argument are
>    compatible.  This is the user's responsibility!
> */
> 
> and I think that should be on the help page.  If it were, this is a
> feature.

Naturally it's a feature!  But I agree that an extra sentance or two on
the help page would be a good idea.  I suggest appending the following
as a new paragraph to the Value section:

As \code{sprintf} is simply a wrapper for the C function of the same
name, it is the user's responsibility to ensure that the objects passed
in \code{\dots} correspond to the format.  Otherwise the resulting
string may contain garbage.  For example, the typical R numeric object
is a float not an integer (see Examples, below).  Likewise, C does not
recognise special values like \code{NA}, so \code{sprintf("Our survey
said: %f", NA)} is not going to work as you might want.

I've cc'ed this to r-bugs.  Jonathan.

-- 
Jonathan Rougier                       Science Laboratories
Department of Mathematical Sciences    South Road
University of Durham                   Durham DH1 3LE
tel: +44 (0)191 374 2361, fax: +44 (0)191 374 7388
http://www.maths.dur.ac.uk/stats/people/jcr/jcr.html

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