[R] gnu ODBC driver for ORACLE in WinXP platform?

Prof Brian Ripley ripley at stats.ox.ac.uk
Wed Oct 7 08:36:57 CEST 2009


On Tue, 6 Oct 2009, Kenneth Roy Cabrera Torres wrote:

> Hi R users and Dr. Uwe Ligges:
>
> I read on the ROracle pre-compiled binary README file that:

Hmm, that is from the CRAN README on Windows pre-compiled binary 
packages: there is no 'ROracle pre-compiled binary' (for Windows or 
Mac OS X) on CRAN.

> "Although the package ROracle passes make check, it seems to be
> dangerous to distribute it:
> I do not have the software available this package depends on."
>
> Why it is dangerous to use ROracle for windows?
> What software does ROracle depends on in windows platform?

Oracle (just like every other platform)!  I believe the statement is 
out-of-date: ROracle seems not to build with current versions of R. 
But the main issues were

(a) It needs to link against Oracle's client software, and that 
depends on the particular version of Oracle.  Bitter experience 
suggests that you need to build such packages from the sources against 
the client software installed on your machine to get reliable service 
from them.

(b) The package does no actual tests in its examples.  It may appear 
to build but not work in practice, since nothing is tested in the 
build process.

> Is there a GNU solution for an ODBC driver on windows XP platform
> for ORACLE for use it with RODBC?

What does GNU have to do with this?  RODBC works with Oracle with 
Oracle's own Windows drivers, and a test suite for Oracle is in the 
sources.  Now, that has been tested against only a couple of versions 
of Oracle, but ODBC is a well-documented API and Oracle's ODBC driver 
writers should always be complying with it.

If perchance by 'GNU' you meant 'GPL' or 'Open Source' then

- Oracle is proprietary, although there are 'free as in beer'
   versions.

- the most open you can be is to communicate with it via an open API,
   and ODBC is such (and I think the only one supported by Oracle).

>
> Thank you for your help.

It does seem churlish to be asking why the hard-working volunteers 
have *not* provided you with a binary version of a package, and such 
queries are more common than thanks for those that *are* provided 
(something like 20x as many).

>
> Kenneth

-- 
Brian D. Ripley,                  ripley at 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 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595




More information about the R-help mailing list