Message-ID: <qsdjfvok0ddss2usevlughuerh67u495g5@4ax.com>
Date: 2003-06-25T15:03:00Z
From: Duncan Murdoch
Subject: excel files and R
In-Reply-To: <Pine.LNX.4.44.0306251149030.2122-100000@gannet.stats>
On Wed, 25 Jun 2003 11:51:44 +0100 (BST), you wrote in message
<Pine.LNX.4.44.0306251149030.2122-100000 at gannet.stats>:
>
>The Excel .xls format is poorly documented, probably deliberately
>obfuscated. A direct interface is on the TODO list: it should be quite
>easy in Windows and possible in other OSes (there is code in Gnumeric, for
>example). In any case, this is a lot harder than the interfaces currently
>in foreign.
The OpenOffice web site has reasonably complete documentation on this
web page: <http://sc.openoffice.org>. It skips some of the more
obscure features, but those probably wouldn't be of interest to R
either.
It looks to me as though it would be relatively easy to write (in R,
using the streams code) a reader that could read strings and numbers.
Interpreting formulas would be a lot harder. (It's possible the last
value of a formula is stored in the .xls file, in which case it would
be easily read too.)
One thing I'm not sure about: recent .xls versions store strings in
Unicode. Does R have cross-platform Unicode support?
Duncan