[R] excel files and R
Marc Schwartz
MSchwartz at medanalytics.com
Wed Jun 25 17:02:14 CEST 2003
On Wed, 2003-06-25 at 09:20, Thomas Lumley wrote:
> On Wed, 25 Jun 2003, Simon Fear wrote:
>
> > I guess all that I and apparently others really want is that "foreign"
> > might
> > include read.excel, like it has read.sas and read.spss. Which is
> > essentially
> > what Bernhard Pfaff's recent post offers - thanks again Bernhard - but
> > using
> > RODBC instead of foreign.
> >
>
> It would be nice, but it's quite hard to read Excel off Windows.
>
> The formats in foreign are either documented by the vendor (accurately in
> the case of Stata and Epi Info, with some omissions for SAS XPORT) or that
> have been reverse-engineered by someone else (read.spss is based on
> PSPP, an attempt at an SPSS clone by Ben Pfaaf, and I think Duncan
> Murdoch did read.S).
>
>
> While it isn't usual to say nice things about commercial vendors on these
> lists I would like to note that Stata not only documents its file format
> in its manuals (with some helpful C snippets for the trickier parts), but
> made available the file format for their `large data set' version 7/SE,
> which I didn't buy.
>
>
> -thomas
Simon,
To add to Thomas' comments and respond to your thoughts, if one were so
inclined, given that R is a volunteer effort, I suspect that an addition
to 'foreign' for Excel would indeed be appreciated by many users.
One resource, with appropriate attribution given, would be the source
code for OpenOffice.org's (OOo) Calc. Since Calc can read and write
Excel formats without using Windows/Office DLL's, it seems reasonable to
presume that OOo has reverse engineered the native Excel file structure.
Since OOo's source is available under the GPL, this could provide the
basis for a "read.excel" function.
Yet another would be Gnumeric, which like Calc is GPL'd and can read and
write native Excel file formats.
More information is available at:
http://www.openoffice.org/dev_docs/source/1.0.3/source.html
http://www.gnome.org/projects/gnumeric/
Food for thought... :-)
Regards,
Marc Schwartz
More information about the R-help
mailing list