Re: [R] Bug/Error in formatC? ....

From: Martin Maechler <maechler_at_stat.math.ethz.ch>
Date: Tue, 10 Jun 2008 16:50:01 +0200

>>>>> "MM" == Martin Maechler <maechler_at_stat.math.ethz.ch> >>>>> on Tue, 10 Jun 2008 15:05:05 +0200 writes:

>>>>> "PD" == Peter Dunn <dunn_at_usq.edu.au> >>>>> on Tue, 10 Jun 2008 09:25:07 +1000 writes:

    PD> Hi all After posting what follows, Duncan Murdoch
    PD> suggested perhaps a bug in formatC, or an error on
    PD> documentation.  Any comments?

I apologize for having slightly misread Peter Dunn's message, and consequently having used the wrong example below.

Indeed, Peter's example

   formatC(0.000059999, digits=2,format="fg",flag="#")

*is* misbehaving in dropping the trailing zero, after having used C-level sprintf() as I indicated below, and I'm currently testing a patch to fix this bug.

Martin Maechler, ETH Zurich.

    MM> It could be called a bug in your platform's
    MM> implementation of the C-library internal sprintf() 
    MM> {to which R's sprintf() is an interface}.
    MM> Things work okay for me on three different Linux platforms and
    MM> on Solaris SPARC.

    MM> More details :

    MM> formatC(x, digits=2, format="fg", flag = "#")

    MM> ends up calling

    MM> sprintf("%#<n>.2g", x)

    MM> where the <n> is carefully determined from x,     MM> in your example

    MM> sprintf("%#3.2g", 0.005999)

    MM> For me, this correctly gives

    >> sprintf("%#3.2g", 0.005999)
    MM> [1] "0.0060"

    MM> and I assume that for your platform, it wrongly returns     MM> "0.006" instead.

    MM> Can you confirm?
    MM> What are the exact platform details?

    MM> Martin Maechler, ETH Zurich

    PD> In particular, bug, error or not, any ideas about how I     PD> can consistently get two significant figures to print?

    PD> P.

    PD> ---------- Original Message ----------

    PD> Hi all

    PD> I am not a C programmer, but I am trying to understand
    PD> formatC to get consistent printing of reals to a given
    PD> number of significant digits.

    PD> Can someone please explain this to me? These first     PD> three give what I expect on reading ?formatC:

>>> formatC(0.0059999, digits=2,format="fg",flag="#")

    PD> [1] "0.0060"

>>> formatC(0.59999, digits=2,format="fg",flag="#")
    PD> [1] "0.60"

>>> formatC(5.9999, digits=2,format="fg",flag="#")
    PD> [1] "6.0"

    PD> This seems consistent with what I read (but perhaps do     PD> not understand) in ?formatC, where I read this:

    PD> digits the desired number of digits after the
    PD> decimal point (format = "f") or significant digits
    PD> (format = "g", = "e" or = "fg").


    PD> Since I am using format="fg" and digits=2, so I am
    PD> expecting two significant digits to always show, which I     PD> have above. So I fail to understand this:

>>> formatC(0.000059999, digits=2,format="fg",flag="#")

    PD> [1] "0.00006"

>>> formatC(0.000059, digits=2, format="fg",flag="#")
    PD> [1] "0.000059"
    PD> I was expecting both of these to produce "0.000059".
    PD> But in the first case above, I get one significant digit
    PD> only.

    PD> I'm obviously misunderstanding something; can someone
    PD> enlighten me? (No doubt, someone will point out a     PD> nuance of the help files I didn't understand!)
    PD> Also, since the above obviously doesn't do what I hoped
    PD> (consistently printing two sig figs), could someone also
    PD> explain how I can consistently get two significant
    PD> figures in situation like above?

    PD> Thanks as always.

    PD> P.

    PD> -- Dr Peter Dunn | dunn <at> usq.edu.au Faculty of
    PD> Sciences, USQ; http://www.sci.usq.edu.au/staff/dunn
    PD> Aust. Centre for Sustainable Catchments:
    PD> www.usq.edu.au/acsc

    MM> ______________________________________________
    MM> R-help_at_r-project.org mailing list
    MM> https://stat.ethz.ch/mailman/listinfo/r-help
    MM> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html     MM> and provide commented, minimal, self-contained, reproducible code.

R-help_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Received on Tue 10 Jun 2008 - 15:10:12 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Tue 10 Jun 2008 - 15:30:39 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive