The basic problem is that by default, RPy converts objects returned
to python to a python object. Unfortunately, this conversion loses
some information, including the type of the original R object.
To avoid this conversion add two lines to your code:
On Feb 22, 2008, at 12:36PM , marco hof
The basic problem is that "survfit" is using deparse() to try to get
the name of the Surv() object so it can generate an R formula.
Unfortunately, this object doesn't have a name in R's namespace, so
deparse() gets the structure representation instead of the name, and
then things fail.
Hello Ding,
It appears that either you don't have numpy properly installed, or
that your copy of rpy was compiled without numpy support.
A simple workaround for this specific case would be to turn of
conversion of r to python objects, so that the object 'x' remains an
r matrix, rather than
This bug should be corrected in the latest version of rpy. What
version are you using?
-G
On Feb 28, 2008, at 10:39AM , Wacek Kusnierczyk wrote:
> Hi,
>
> Having used rpy for just a while on Fedora 7, python 2.5, I discover
> that it crashes on graphics functions such as plot, hist, etc.
>
>
Wow, Laurent, you've been doing a lot of work!
On Mar 2, 2008, at 2:52PM , <[EMAIL PROTECTED]>
<[EMAIL PROTECTED]> wrote:
> (there is something odd going on, with variable names
> at the Python level suddenly pointing to nowhere - and
> this often ends in a segfault :/ )
Probably R is garbage c
You can pass the read data directly to R. Probably the best way to
do this is to construct a python list containing one vector per
column. Call as.data.frame() on the list, and then proceed.
For example:
from rpy import *
set_default_mode(NO_CONVERSION)
data = [ [
Hi All,
I think it is a good idea to use a Wiki to help develop
documentation. I've enabled Wiki support over at sourceforge for rpy at
http://rpy.wiki.sourceforge.net/
but I don't have time to try to pull the existing documentation in
there.
Of course, patches against the existin
If you explicitly set the RHOME environment variable to point to the
R 2.4.1 directory, rpy will use that RHOME and hence R 2.4.1.
-Greg
On Apr 10, 2008, at 5:53AM , Eloi Ribeiro wrote:
Dear list,
We are facing the fowling problem. Every thing worked fine while we
had Python 2.4 + R 2.
Hi All,
I'm sorry to hear that the C stack limit issue has arisen again.
I'll check the source code to see if the fix that we put in for RC3
has gotten removed or disabled somehow.
-Greg
On Apr 9, 2008, at 2:44PM , Laurent Gautier wrote:
I can reproduce it with python-2.5 / R-2.6.1 as
.
-G
On Apr 10, 2008, at 9:17AM , Gregory Warnes wrote:
Hi All,
I'm sorry to hear that the C stack limit issue has arisen again.
I'll check the source code to see if the fix that we put in for RC3
has gotten removed or disabled somehow.
-Greg
On Apr 9, 2008, at 2:44PM , Laure
Hi Mark
I do my rpy development on OSX 10.4, using python 2.4.2.
To install rpy, just download and unpack the source .tar.gz file,
then run
python setup.py build
sudo python setup.py install
you'll probably want to ensure that Numpy is installed before you do
this so you
Hello Andrew,
The basic problem you are encountering is that python uses a
dictionary to store named parameters to function calls. Python's
dictionaries don't preserve order, just name-value correspondence.
As a consequence calling an R function with named arguments loses the
ordering.
The fix for this bug has been submitted to SVN.
-G
On May 12, 2008, at 2:26PM , John Reid wrote:
Well I seemed to have fixed it by changing the call to
KillAllDevices()
in rpymodule.c to Rf_KillAllDevices() as is defined in libR.so.
Perhaps there's a better way?
Thanks for a great rpy pac
The standard Mac version of R includes the necessary shared library,
hence it is not necesary to do this. You should be able to simply run
cd /path/to/rpy/source/code/
python setup.py build
sudo python setup.py install
and everything should install properly, at lea
Hi Ernesto,
These warning messages are normal, albeit annoying.I've just made
commited changes to the the SVN repository to avoid the ones from
rpymodule.c and io.c, so this should be less of a problem in the future.
-Greg
On May 14, 2008, at 3:34AM , Ernesto wrote:
Dear all,
I'm
Too bad.
I guess I'll have to either add code to look for
libRlapack.so
in the appropriate directory. Know of any simple way to do this?
-G
On May 14, 2008, at 10:16AM , Dirk Eddelbuettel wrote:
Hi Greg,
On Wed, May 14, 2008 at 10:08:29AM -0400, Gregory. R. Warnes wrote:
>
> Hi
I just committed a change to use R CMD config LAPACK_LIBS. Can you
check if it behaves properly?
-G
On May 14, 2008, at 1:14PM , Dirk Eddelbuettel wrote:
On Wed, May 14, 2008 at 01:06:38PM -0400, Gregory Warnes wrote:
> Too bad.
>
> I guess I'll have to either add co
Yea!
On May 14, 2008, at 3:08PM , Dirk Eddelbuettel wrote:
On Wed, May 14, 2008 at 02:36:27PM -0400, Gregory Warnes wrote:
> I just committed a change to use R CMD config LAPACK_LIBS. Can you
> check if it behaves properly?
I get the commit emails, so I just hand-applied the changes
Hmm.
Does rerunning the install make a difference?
-G
On May 14, 2008, at 4:44PM , Manu Hack wrote:
Hi,
I just installed R and rpy on Opensuse 10.3. I managed to get it
work once. But then all of a sudden it won't work again. Here is
the info.
I compiled R-2.7.0 with configure --en
Is there a difference in the environment variables, particularly the
value of RHOME?
-G
On May 14, 2008, at 5:39PM , Manu Hack wrote:
On Wed, May 14, 2008 at 4:49 PM, Gregory Warnes
<[EMAIL PROTECTED]> wrote:
Hmm.
Does rerunning the install make a difference?
-G
No, on
Hello Lars,
You will need to create an R function to accomplish the needed task.
One way is to put everything into one bug R function:
from rpy import r
r.pdf(filename,onefile=r.TRUE, width = 7.8, height = 7,)
my_contour = r("""
function(x, y, z, n=20)
{
Lab.palette <- colorR
.
-G
On May 14, 2008, at 6:08PM , Manu Hack wrote:
On Wed, May 14, 2008 at 5:50 PM, Gregory Warnes
<[EMAIL PROTECTED]> wrote:
Is there a difference in the environment variables, particularly
the value of RHOME?
-G
No, RHOME are the same.
On May 14, 2008, at 5:39PM , Manu H
The best way to store R objects is to use R's own 'save()' and 'load
()' functions. In order to use these, it is necessary for the
objects to be assigned to R names, so you would have something like...
from rpy import r
# code here to create objects 'a' and 'b'
r.as
I think this is probably a Numpy installation issue. What happens if
you issue the command
import numpy
??
-Greg
On May 22, 2008, at 1:40PM , James D Dickinson wrote:
Hi - I'm fairly new to Python and I'm having a heck of a a getting the
install of 'rPy' to go.
I downloaded th
Hmm. How about the more adventuresome
import numpy.core.multiarray
?
-Greg
On May 22, 2008, at 3:59PM , James D Dickinson wrote:
>>> import numpy
works just fine no messages or errors -
trying to import rpy results in same messages below
-James
>I think this is prob
OK, this is quite odd. Two more things to try. First, try adding
the directory containing multiarray.DLL to the PATH before running
python. Second, if this doesn't work, try running python with the -v
option and redirecting the output to a file..
-G
On May 22, 2008, at 4:21PM , James
Hi Laurent,
Why not use the get_R_HOME() function in rpy_tools.py, which already
handles this for Unix and Win32. (On Win32 it checks for the proper
R registry entry for this information, then falls back to other
methods if this is not available.)
-G
On Jun 1, 2008, at 6:54AM , <[EMAI
Makes good sense.
-G
On Jun 1, 2008, at 7:40AM , <[EMAIL PROTECTED]>
<[EMAIL PROTECTED]> wrote:
Revision: 541
http://rpy.svn.sourceforge.net/rpy/?rev=541&view=rev
Author: lgautier
Date: 2008-06-01 04:40:33 -0700 (Sun, 01 Jun 2008)
Log Message:
---
- only keep suppo
Aside, since rpy _does_ check for specific R versions, would it not
be advisable for the .deb dependencies to reflect this appropriately?
-G
On Jun 8, 2008, at 1:13PM , Dirk Eddelbuettel wrote:
On 8 June 2008 at 11:15, Lars Tangvald wrote:
| Laurent Gautier skrev:
| > 2008/6/7 Dirk Edde
You may find the forumula syntax
y ~ .
helpful. If data= is supplied to the modeling command, the single
period means 'all variables in the data frame that are not
otherwised mentioned.
As for changing he names on an existing object, you can't use the
rownames(x) <- foo syntax
Thanks Peter, the sf.net news page has been disabled.
On Jun 9, 2008, at 10:05AM , Peter wrote:
I'm confused about why there are two different news pages:
http://rpy.sourceforge.net/news.html
This includes the new wiki announcement and the 1.0.2 release and a
lot of historical entries. Thi
versions.
-G
On Jun 9, 2008, at 1:12PM , Dirk Eddelbuettel wrote:
On Mon, Jun 09, 2008 at 12:36:19PM -0400, Gregory Warnes wrote:
> Aside, since rpy _does_ check for specific R versions, would it
not be
> advisable for the .deb dependencies to reflect this appropriately?
I tried that
On Jun 10, 2008, at 9:20AM , Chris Arthur wrote:
Thanks Gregory, that "y ~ ." syntax solves my problems. and allows
my to
pull out the PLSR data I wanted, but leads me to my next problem.
Glad to help.
I then try to use my model to predict the value of the matrix accFAS
predicted=r.pr
The Cairo package is an optional library. You should be able to install it
using:
> install.packages(Cairo)
And then get information on using it via:
> library(Cairo)
> ? Cairo
-Greg
On 6/16/08 4:13PM , "Patrick Jackson" <[EMAIL PROTECTED]> wrote:
> I believe we are finally getting somewher
Hi Paul,
We could be of more assistance if you included some source code...
Are you doing the looping in R or in python?
-Greg
On 6/23/08 10:32AM , "Jung, Paul" <[EMAIL PROTECTED]> wrote:
> I'm having the following issue with rpy. I'm calling r.lm to do a bunch
> of regressions (several hun
On 6/24/08 10:52PM , "Nishant Joshi" <[EMAIL PROTECTED]> wrote:
> Referenced from: /Library/Python/2.5/site-packages/libR.dylib
Where exactly are your libR.dylib, libRblas.dylib, and libRlapack.dylib
files? This doesn¹t look like the standard location...
The standard locations on the mac is:
Hello Laurent,
The R system itself is not thread safe, and has quite a bit of persistent
state. Therefore if you want to use R from multiple threads, you will need
to arrange to have a single thread interact with R at a time via appropriate
locking or delegation.
-G
On 6/26/08 11:17AM , "lau
Side note: I developed the RSOAP package to allow a single multi-threaded
python application to interact with multiple independent R processes
cleanly. Take a look at http://rsoap.sourceforge.net or
http://sourceforge.net/projects/rsoap/
-Greg
On 6/26/08 11:17AM , "laurent oget" <[EMAIL PROTE
gt; >> 2008/6/26 laurent oget <[EMAIL PROTECTED]>:
>>>> >>> So if I want to take advantage of my two cores, i need two processes,
>>>> right?
>>>> >>>
>>>> >>> 2008/6/26 Gregory Warnes <[EMAIL PROTECTED]>:
&
uments)
> complicated (if possible).
>
> Ideas, or help, on the matter are welcome.
>
>
> L.
>
>
>
>
> 2008/6/26 Gregory Warnes <[EMAIL PROTECTED]>:
>> >
>> > Hello Laurent,
>> >
>> > The R system itself is not thr
of rpy
> (may be a class extending Python's "threading.Thread" ?)
>
This makes good sense.
-G
>
> L.
>
>
>
>
> 2008/6/26 Gregory Warnes <[EMAIL PROTECTED]>:
>> > Hi Laurent,
>> >
>> > It does seem desirable to inc
Hi Jacob,
It looks like the event handler code isn't properly setting or
resetting the control-c handler. Can you try seeing whether Control
C is handled correctly while R itself is processing. IE, exectute:
from rpy import r
r("for(i in 1:1e6) 1+1")
then press Control-C.
s with my small samples to produce the
> issue more frequently.
>
> I realize this doesn't resolve your question with more than a
> "sometimes". Is there a case that might test this more definitively?
>
> -Jacob
>
> On 03/26/2007 03:21 PM, Gregory Warnes wrote:
>>
doesn't resolve your question with more than a
> "sometimes". Is there a case that might test this more definitively?
>
> -Jacob
>
> On 03/26/2007 03:21 PM, Gregory Warnes wrote:
>> Hi Jacob,
>>
>> It looks like the event handler code isn't
ible actions", keyboard input
> goes
> to python or nowhere, but never the menu itself.
>
> Thanks again.
> -Jacob
>
> On 03/26/2007 04:14 PM, Gregory Warnes wrote:
>> Hmm. Try this code instead, maybe it'll be slow enough to ensure you
>> ca
Thanks for letting me know about the broken links, I'll see about
getting them fixed.
-G
On Apr 3, 2007, at 4:08AM , Dumoulin Nicolas wrote:
> Hi there,
>
> I've just discover the rpy project, it looks great !
>
> Before to ask technical questions (With a so beautiful manual,
> maybe I will
Done. The mailing list archive and Tim Church's demo are now
accessible again.
-G
On Apr 3, 2007, at 10:07AM , Gregory Warnes wrote:
> Thanks for letting me know about the broken links, I'll see about
> getting them fixed.
>
> -G
>
> On Apr 3, 2007, at 4:0
Hi All,
The missing files have been added to the site.
-G
On Apr 16, 2007, at 11:58AM , Alexander Belopolsky wrote:
> On 4/16/07, Fabrice Tiercelin <[EMAIL PROTECTED]> wrote:
>
>> faithful.dat - the link is broken
>
> I guess Gregory will have to fix the links. Meanwhile you can get the
> files
Hello Erez,
On Apr 17, 2007, at 3:56PM , Erez Lieberman wrote:
> Hey All -
>
> I'd really love to get some insight into this, I'm just totally
> puzzled!
>
> I'm trying to make system calls to clustalw, a sequence aligner
> which takes arguments like filenames. The main point is just that I
Hi Sean,
RPy should be perfectly happy with S4 classes, so you should be able
to do what you want.
-G
On Apr 21, 2007, at 7:20PM , Sean Davis wrote:
> I haven't started playing much with S4 classes from RPy. I have
> used S4 classes a good bit in R, but wondered if anyone had some
> poi
Hi John,
If the gcc's are incompatible, then you may have yet another
problem: python itself will have been built with one, and R the other.
Before trying to modify the default compiler options, try running the
compler commands that setup.py displays as it tries to work. It
should pick up
Sorry, there was a typo in my message:
On May 13, 2007, at 10:05PM , Gregory R. Warnes wrote:
>
> The problem is that in **R** '.' (period) is a valid character in a
> variable/function *name*, while it is not valid in python. To
> overcome this problem, use '_' (underscore), which RPy will tran
Hello John,
The correct syntax for this in R is:
axis(1, at=1:5, labels=expression(1^10, 2^10, 3^10, 4^10, 5^10))
Note that expression() expects actual R syntax and not strings. To
handle strings you would need to do
axis(1, at=1:5, labels=parse(text=c("1^10", "2^10", "3^1
Hi Pierre,
The easiest way to keep the window open is to prevent the script from
exiting until you are done with the display. This can be
accomplished by something like:
r.readline("Press to exit: ")
-Greg
On Jun 7, 2007, at 10:24AM , pierre ratinaud wrote:
> I all,
> sorry if this is
Hi Pierre,
I'm sorry to hear that readlines() doesn't work properly under
wxPython. It sounds like I need to take a look at it. In any case,
The important concept is to prevent the script from exiting until you
are done with the graph. There are many ways to accomplish this.
-G
On Jun
Have you verified that this package will load correctly directly in R?
-G
On Jun 13, 2007, at 12:05AM , Антонец Денис wrote:
> Do you know the reason why RPy failed to load "kernlab" package?
> Other packages were loaded correctly... Could this be explained by the
> fact that "kernlab" uses S
Hello,
I've loaded and run both e1071 and kernlab on my system (Mac OSX)
using rpy. Please provide some information about your system so we
can figure out what the problem is.
Operating System
Python Version
R Version
RPy Version
Output of a complete session:
from rpy import r
The normal way to make the call to an r function using rpy is to call
it like so:
import rpy
rpy.r.function_name(a=1, b=2, c=4)
note that in R, the period character '.' is a valid part of names,
while in python it is not. To overcome this problem, rpy translates
the under
Hi Michael,
The trick here is that r['...'] expects the ... to be an object name,
rather than a statement to be executed. The brackets following the
name 'r' is shorthand for 'obtain the object named ... from r', which
can also be expressed as,
r.get('...')
RPy provides a similar s
Hello Mohar,
This question is outside of the RPy scope. It looks like the type of
question that should be posed to the R-devel mailing list:
[EMAIL PROTECTED]
-Greg
On Jun 20, 2007, at 8:25PM , Mohar Chattopadhyay wrote:
Dear all,
I have encountered errors in compiling R-2.5.0
Hi Tomi,
I know it has been a while, did you get the help you needed?
-Greg
On May 4, 2007, at 12:22PM , Tomi Korhonen wrote:
> Hello,
>
> I'm trying to use Rpy in a MPI-environment ( pypar ) where I
> distribute my
> processing to multiple CPU's. I'd like to be able to plot out the
> resu
Hello Gordan,
Did you find a solution to your problem?
-G
On May 25, 2007, at 8:06AM , Gordon Bergling wrote:
> Hi folks,
>
> I am currently trying to create a port of an linux application thats
> uses python/pygtk/rpy. On my WinXP installation I have installed
> R-2.5.0, python-2.5 and the ne
Hi All,
I've recently launched a software company, Random Technologies LLC
that provides commercial packaging, support, services and validation
materials for R, RPy, and RSOAP.
For more information, please visit our web site:
http://www.random-technologies-llc.com
review our flyers
Hi Skip,
Yes, it is necessary to build R as a shared library in order for RPy
to function. Have you checked for information / asked for assistance
on the R-help list for this problem?
BTW, what platform are you trying to build on?
-Greg
On Jul 2, 2007, at 4:53PM , Skip Montanaro wrote:
>
Hi Mark,
I regularly use rpy on my MacBookPro. What compiler's are you using?
(Ie
gcc -v
)
-G
On Jul 31, 2007, at 3:34PM , Mark Thomas wrote:
> Hi
>
> I have spent most of today trying to install rpy on my Mac
> (osx10.4.10). I
> have python 2.5.1 and R version 2.5.0 (2007-04-23)
Hello Manuzhai (and RPy'ers)
I try to be pretty responsive, but have been really busy the last
month or so setting up a new company to provide commercial packaging,
support and services for R and RPy. Check us out at http://random-
technologies-llc.com.
I'll take a look at your patches, an
Hi David,
You need to use R's print function, rather than python's:
from rpy import *
data = {
"y":[1,2,3,4],
"x":[1,1,2,2]
}
set_default_mode(NO_CONVERSION)
lm = r.lm(r("y~x"),data)
summary_lm = r.summary(lm)
tmp = r.print_ (summary_lm)
Note the u
To use Rpy to compute the standard deviation to something like:
>>> import rpy
RHOME= /Library/Frameworks/R.framework/Resources
RVERSION= 2.5.1
RVER= 2051
RUSER= /Users/warnes
Loading Rpy version 2051 .. Done.
Creating the R object 'r' .. Done
>>> x = [1,2,3,4,5]
>>> rpy.r.sd(x)
1.58113883008
Something seems to be awry with Python 2.5. Can you check Python 2.4?
-G
On Oct 8, 2007, at 5:18PM , Vince Fulco wrote:
> A review of the docs indicates that conversion s/b automatic. I am
> trying to do some elementary calculations on a numpy array and am
> getting the following error.
>
>
It sounds like your version of RPy may have been compiled without
NumPy support. Did you compile it yourself, or are you using a pre-
build version.
-G
On Oct 8, 2007, at 5:18PM , Vince Fulco wrote:
> A review of the docs indicates that conversion s/b automatic. I am
> trying to do some e
Check out http://research.warnes.net/projects/RStatServer for a
complete Zope+Rpy+R solution.
-G
On Aug 28, 2007, at 8:13AM , Rémi Flicoteaux wrote:
> Hello,
>
> I trying to use R from zope / plone server using rpy interface with
> python script on the zope server.
>
> Some errors occurred: w
I've been trying to make time to work on fixing all of the
outstanding bugs. Other duties keep pulling me away.
This should be a simple fix...
On Oct 16, 2007, at 7:56PM , Toby Dylan Hocking wrote:
> Hi all,
>
> I really need RPy to resolve this issue:
> http://sourceforge.net/tracker/index
to use that.
>
> I'm going to take a short break from this for a bit, but please
> email me any suggestions/ideas and what to do next.
>
> Toby Dylan Hocking
>
> -Original Message-
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] Behalf Of Gregory
> War
I'm in the process of bug-fixing and patching. I expect to post a
new source code release early next week, and will try to follow up
with a windows build soon after.
-G
On Oct 17, 2007, at 6:58PM , Mario Beauchamp wrote:
> Hi all,
>
> I am having the same probs mentioned previously import
Hi Alex,
I fixed a couple of Windows startup bugs and made the commit
yesterday evening. So please try using the latest SVN source.
-G
On Oct 18, 2007, at 11:31AM , Alex Mandel wrote:
> Peter wrote:
>> Alex Mandel wrote:
>>> A quick questions on getting started with Rpy:
>>>
>>> 1. Does t
Hi All,
I've made a bunch of changes in the SVN source code repository, some
of which are specifically related to Numeric/NumPy conversion. Can
y'all give the new code a try to see if it helps?
-G
On Oct 24, 2007, at 6:04PM , Matthew Brett wrote:
>> I can reproduce the problem, yes.
>>
>>
Thanks, I'll take a look when I'm off the clock!
-G
On Oct 25, 2007, at 9:17AM , Matthew Brett wrote:
> Hi Gregg,
>
>> I've made a bunch of changes in the SVN source code repository, some
>> of which are specifically related to Numeric/NumPy conversion. Can
>> y'all give the new code a try to
I'm off on a consulting project this week, but will be testing out
the latest patch and (hopefully) posting a new release (which will be
labeled 1.0.0) this evening, all going well.
-G
On Oct 25, 2007, at 12:18PM , Alex Mandel wrote:
> Skip Montanaro wrote:
>> Does Rpy work with R 2.5.1
On Oct 25, 2007, at 12:50PM , Skip Montanaro wrote:
> Alex & G(reg(ory)?)?,
LOL!
I use Gregory for formal stuff, Greg for my friends, and sign email
with -G. ;-D
-G
-
This SF.net email is sponsored by: Splunk Inc.
Stil
These messages are generated by the R shared library as it attempts
to load the fundimental R packages it needs. This usually means
that your installation of R is lacking properly installed versions of
these packages. Does R start up and run properly when run directly?
-G
On Nov 2, 200
ally use R independently of rpy, so I'm not
sure of how to test that they are working correctly other than
that. Is there something I can run to test it?
On 11/5/07, Gregory Warnes <[EMAIL PROTECTED]> wrote:
These messages are generated by the R shared library as it attempts
to load the f
Hi JL,
I'm in the proccess of fixing a number of bugs with RPy, and will be
releasing a new version shortly that resolves this issue. If you
are impatient, you can get the current version from the subversion
repository, where the fix for this issue has already been committed.
-G
Nov 7,
Hi Bo,
As soon as I confirm that all the tests run properly, I'll post the
source package for 1.0.1. I'll be glad to let you build and upload
the windows binary version...
-G
On Nov 29, 2007, at 11:48AM , Bo Peng wrote:
>> Somehow I missed the removal of the 'xstr(...)' when I generated
Ok, first, on windows, rpy attempts to add the necessary locations to
the PATH environment variable so that DLLs are properly found. If
the new version doesn't solve the problem, a work-around is to
manually set up the path before starting python (or from within
python) so that the direc
OK, RPy 1.0.1 with the windows fix is up, including binaries. Please
take them for a spin and let me know if there are any problems.
-G
On Nov 29, 2007, at 1:33PM , Gregory Warnes wrote:
> Hi Bo,
>
> As soon as I confirm that all the tests run properly, I'll post the
> s
Hi Etienne,
The basic problem is that under the default conversion mode
(BASIC_CONVERSION) all R objects are converted to roughly-equivalent
python structures. As a consequence, the object 'av' isn't actually
an R object, so r.summary(av) won't treat it as such. The simplest
solution i
n R-->Python happens,
some information is lost, so the av object itself isn't a valid aov R
object when Python--R conversion happens.
-G
> However, I tried to call explicitly summary_aov() with no success.
> But I have to test it before to go further.
> Thanks again !!
> -Etie
When asking a question of this sort, it is usually helpful to provide
an example of the code you were attempting to run.
For what its worth, the scale function works properly for me:
>>> from rpy import *
>>> r.scale( range(10) )
array([[-1.48630108],
[-1.15601195],
[-0.82572
Hi Mario,
Can you check the python local packages directory to check the
version of rpy that is actually getting installed. I did have a
small problem in the first version of the file I uploaded, but I
thought that I had fixed it
-G
On Dec 19, 2007, at 10:04PM , Mario Beauchamp wrote
Hi Barry,
I suspect the problem is that QGIS has set up python's console input
differently than rpy expects. At the moment, there is no way to
tell RPy to do something different with i/o during startup. (After
startup you can call set_rpy_output() and set_rpy_input() to provide
new i/o
Hi Laurent,
I'm glad to see your contributions. Robj, in particular, has needed
to be expanded to support all of the (appropriate) python operations
for some time. Make sure and look at the stuff in /sandbox ...
Since 2.2 was released Dec 2001, I expect that we can now drop
support for p
Hi Barry,
Grab the latest SVN version of rpy. I've added the ability to avoid
setting up reading and writing the console via rpy_options:
import rpy_options
rpy_options.set_options(SETUP_READ_CONSOLE=False,
SETUP_WRITE_CONSOLE=False, SETUP_SHOWFILES=False)
from rpy im
> On Dec 20, 2007 5:20 PM, Gregory Warnes <[EMAIL PROTECTED]>
> wrote:
>> Hi Mario,
>>
>> Can you check the python local packages directory to check the
>> version of rpy that is actually getting installed. I did have a
>> small problem in the first versio
> in touch with him to know what is in there ?
>
> Laurent
>
>
> 2008/1/3, Gregory Warnes <[EMAIL PROTECTED]>:
>>
>> Hi Laurent,
>>
>> I'm glad to see your contributions. Robj, in particular, has needed
>> to be expanded to support all of th
Hi Mario,
> I had stepped through rpy.py in pythonwin and everything was normal.
> It crashes pythonwin in the exec("import _rpy%s as _rpy" % RVER)
> statement, and RVER
> is the right one.
>
Oh dear.
> I will investigate further tomorrow, when I'm on my XP box.
>
Ok, thanks.
-G
> cheers
>
On Jan 7, 2008, at 11:32AM , Barry Rowlingson wrote:
> Gregory Warnes wrote:
>>> Thanks for the help, not sure where to go now except maybe if
>>> rpy had an option to pass --no-Rconsole to R on startup...
>> Rpy can pass anything we want to R, but as far as I can
Oh, thank goodness. I'm glad it was an easy fix. :-D
-G
On Jan 7, 2008, at 3:01PM , Mario Beauchamp wrote:
> Hi Gregory,
>
> On Jan 7, 2008 10:16 AM, Gregory Warnes <[EMAIL PROTECTED]>
> wrote:
>>
>> Hi Mario,
>>
>>> I had stepped through rpy
There was also a bug in RPy's handling of unicode strings until 1.0,
so make sure you have a recent version of RPy as well.
-G
On Jan 8, 2008, at 2:47AM , Laurent Gautier wrote:
> Crashes with unicode labels in plots was reported as a problem
> with R-2.3 to R-2.5.
> It appears fixed with th
Thats odd. If you've got the setup to compile RPy, you might try
adding an explicit "--no-Rconsole" argument to the variable
"defaultargv" at the top of src/rpymodule.c and see if that makes a
difference for the 'out of the box' R.
-G
On Jan 8, 2008, at 12:11PM , Barry Rowlingson wrote:
R itself is neither multithreaded, nor thread safe. If you want to
interact with more than one R session, you'll have to have a separate
process for each R session. I built the RSOAP tool, which includes
wrapper (called something like) LocalRSOAPConnection that wraps the
separate session
1 - 100 of 102 matches
Mail list logo
