Thanks Laurent.
I have installed a VMWare Ubuntu machine and am quite impressed.
Python runs much faster as a Linux VM under Win7 than it did on a dual
boot XP (quad core I7).
However I have been having no end of problems trying to get rpy2 to run.
When I run the tests

python -m 'rpy2tests'

I get the following error

  File "/usr/local/lib/python2.6/dist-packages/rpy2/rinterface/__init__.py",
line 75, in <module>
    from rpy2.rinterface.rinterface import *
ImportError: libR.so: cannot open shared object file: No such file or directory

I took great pains to compile the R11.1 with the following

./configure --enable-R-shlib
make
make install

which passed OK.
What else could be wrong?

Andrew.


On Sun, Aug 8, 2010 at 2:54 AM, Laurent Gautier <lgaut...@gmail.com> wrote:
> Not necessarily. rpy2-2.0.8 can mostly do what you like, however may be not
> as simply / optimally as the rpy2-2.1.x series could.
>
> My recommendation would be to go for the 2.1.x series, or the 2.2.0-dev
> series if you are of the adventurous kind.
>
> There is also more documentation for those, more examples, and optimization
> hints (which when followed are believed to make rpy2 >= 2.1.0 the most
> efficient and versatile Python-to-R bridge).
>
> 2.1.x and 2.2-dev series are currently working fine on POSIX systems (Linux,
> BSD, etc...), but Windows development is currently halted until there is
> support for it. In the meanwhile, you could consider installing a linux on a
> virtual machine ("virtualbox", or a paying solution if you have the
> resources for it - "Parallels" for example).
> If your application must also interface with Windows components, you could
> consider a client-server approach.
>
> L.
>
> PS: May be you meant "factanal" instead of "factoran" ?
>
>
> On 8/7/10 5:58 PM, Andrew Walsh wrote:
>>
>> Yes thanks Laurent.
>> So if I want to use factoran from R stats package I presume this means
>> my only option is to install Linux somewhere rather than use windows
>> (if I want to use it from Rpy2 which I do).
>>
>> Andrew.
>>
>> On Sat, Aug 7, 2010 at 9:38 PM, Laurent Gautier<lgaut...@gmail.com>
>>  wrote:
>>>
>>> So we'll all assume that you both found out that:
>>>
>>> - rpy2.robjects.packages only exists for rpy2>= 2.1.0
>>>
>>> - the windows binaries are currently stuck in the past, 2.0.8 being the
>>> latest version available.
>>>
>>>
>>>
>>> On 8/7/10 12:05 PM, Andrew Walsh wrote:
>>>>
>>>> Huge apologies everyone. I thought I looked through the list but
>>>> obviously not very well at all.
>>>>
>>>> On Sat, Aug 7, 2010 at 8:03 PM, Andrew Walsh<tsar...@gmail.com>
>>>>  wrote:
>>>>>
>>>>> I have installed Rpy2 v 2.08 under Windows XP 32. I want to do a
>>>>> factor analysis from the R stats package.
>>>>>  From the RPY2 documentation I need to import rpy2 packages. I find I
>>>>> can import rpy2 and rpy2.objects but rpy2.robjects.packages cannot be
>>>>> found.
>>>>>
>>>>>
>>>>>>>> import rpy2
>>>>>>>> import rpy2.robjects
>>>>>>>> import rpy2.robjects.packages
>>>>>
>>>>> Traceback (most recent call last):
>>>>>  File "<pyshell#2>", line 1, in<module>
>>>>>    import rpy2.robjects.packages
>>>>> ImportError: No module named packages
>>>>>
>>>>> Any suggestions please?
>>>>>
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> This SF.net email is sponsored by
>>>>
>>>> Make an app they can't live without
>>>> Enter the BlackBerry Developer Challenge
>>>> http://p.sf.net/sfu/RIM-dev2dev
>>>> _______________________________________________
>>>> rpy-list mailing list
>>>> rpy-list@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/rpy-list
>>>
>>>
>
>

------------------------------------------------------------------------------
This SF.net email is sponsored by 

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
rpy-list mailing list
rpy-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rpy-list

Reply via email to