[R] ReadLines question

[Roger Bivand]

On 21 Oct 2006, Peter Dalgaard wrote:

> Roger Bivand <Roger.Bivand at nhh.no> writes:
> 
> > On Sat, 21 Oct 2006, Jonathan Greenberg wrote:
> > 
> > > That looks to me like an infinity sign (I have no idea why that is part of
> > > the header of this file, but it is there).  How do I modify the encoding to
> > > read this in? 
> > 
> > The problem is the degree sign. Under linux:
> > 
> > $ file tmp/Marlette_lake_snotel.csv
> > tmp/Marlette_lake_snotel.csv: ISO-8859 text, with CRLF, CR line terminators
> > 
> > so probably the conversion to multibyte is happening on your reading 
> > platform. Reading the file into 2.4.0 on Windows with a Norwegian 1252 
> > setting (Sys.getlocale()), I see the degree sign.
> 
> Nono, there is no conversion. R is _expecting_ a multibyte sequence
> (utf8 most likely) and finding something that isn't part of one.  
> 
> The fix should be something close to
> 
> read.csv(file("tmp/Marlette_lake_snotel.csv"), encoding="iso-8859-1") 
> 

Yes, thanks:

read.csv(file("tmp/Marlette_lake_snotel.csv", encoding="iso-8859-1"), 
 skip=1, check.names=FALSE) 

gives usable results on Linux/en_GB and Win/Norwegian (Bokmål) 1252.


> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no