According to the help page, freq and prob are logical, so this is
correctly an error.
What it actually does is
freq <- if (!missing(probability)) !as.logical(probability)
which is consistent with the documentation.
On Thu, 10 Feb 2005 tomhopper@comcast.net wrote:
> Full_Name: Tom Hopper
> Version: 2.0.1
> OS: Windows 2000
> Submission from: (NULL) (69.220.229.2)
>
>
> While attempting to wrap hist() in my own custom function (for formatting
> purposes), I notice that a call like
>
> > hist(x, prob=NULL)
>
> produces the output:
>
> Error in if (freq) x$counts else { : argument is of length zero
>
> Switching to use freq does not produce an error, even when freq = NULL.
>
> It appears that hist() is checking for the condition
>
> !missing(prob)
>
> but needs to add a check for a NULL value
>
> && !is.null(prob)
--
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 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595