[R] Excel *.xls files, RODBC

[Dirk Eddelbuettel]

On Sun, Dec 05, 2004 at 12:58:19PM +0000, Gabor Grothendieck wrote:
> Rolf Turner <rolf <at> math.unb.ca> writes:
> 
> : 
> : I gather from reading the back-issues of r-help that it should be
> : possible (modulo a number of caveats) to read an excel (yuck!) file
> : into R using RODBC.  I have obtained and installed ODBC and the RODBC
> : package, but cannot for the life of me figure out how to go about
> : it.  Can anyone give me a simple recipe?
> : 
> : I have an excel file on cdrom, say:
> : 
> : 	/mnt/cdrom/melvin.xls
> : 
> : I have started R and loaded the RODBC package.  I want to create
> : a data frame ``melvin'' by reading in /mnt/cdrom/melvin.xls.
> : What (in monosyllables --- step by step) do I do next?
> 
> The xlhtml program at
> 
>   http://freshmeat.net/projects/xlhtml/
> 
> can not only convert .xls to .html but also to .csv using something
> like:
> 
>   xlhtml -te -xc:1-10 -csv

AFAIK there are about three main strands of tools to deal with this in a
manner that is platform-independent:

-- Perl based on SpreadSheet::ReadExcel and OLE::StorageLite, which Greg has
   wrapped up nicely in gdata, a component of the gregmisc bundle
   
-- C based libraries also used in Gnumeric, and, for that matter, also as a
   loadable module for GNU Gretl -- someone ambitious could add this to the
   foreign package; this may make for a nice term project. OpenOffice may have
   its own code base.
   
-- Apache/Jakarta/POI based, which I know little of, but Whit just told us
   that he has something in the works there

I think all three of these are maintained (in fact, I look after the Perl
and Gretl ones for Debian).  Where does xlhtml fit in?  The code seems to be
C based -- it this a split of the Gnumeric code?  Have there been updates
since 2002?

Dirk

-- 
If you don't go with R now, you will someday.
  -- David Kane on r-sig-finance, 30 Nov 2004