[Rd] problem with load('http://....') (PR#2344)

[ripley@stats.ox.ac.uk]

On 4 Dec 2002, Peter Dalgaard BSA wrote:

> fharrell@virginia.edu writes:
>
> > Full_Name: Frank Harrell
> > Version: 1.6.1
> > OS: RedHat 8.0 Linux
> > Submission from: (NULL) (128.143.108.90)
> >
> >
> > I get an error when trying to load a URL that contains a file that was saved
> > using save(object, compress=TRUE):
> >
> > > load('http://hesweb1.med.virginia.edu/biostat/s/data/sav/kprats.sav')
> > Error in gzfile(file, "rb") : unable to open connection
> > In addition: Warning message:
> > cannot open compressed file
> > `http://hesweb1.med.virginia.edu/biostat/s/data/sav/kprats.sav'
> >
> > The above file will remain at that URL for your testing.
>
> The root of the issue is that gzfile() doesn't know about URLs. Not
> documented to either, but (on the face of things) it would seem easy
> to add the feature.

Right, this is not what the `file' argument of load() is documented as.
That is neither a file name nor a connection.

I am worried about attempting to get a binary file over a url connection,
but in principle what you should do is

load(url('http://hesweb1.med.virginia.edu/biostat/s/data/sav/kprats.sav'))

That fails I think because the connection cannot handle binary files,
but I will look further when I have time.

-- 
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