I realize this has been hammered to death, but I can't seem to get my head
around it.
Is the mapping from rpy (or rpy2) to R a one-to-one? Can either invoke
multiple R processes from a single python program?
In the past I've used python threads and piped stdin/stdout to multiple R
processes to t
Thanks Nathaniel!
> Spawn multiple Python processes, each of which uses rpy2 to load
R, and talk to them over any Python-level IPC library.
The latter is really a huge improvement over the former, because there
are great IPC libraries for python -- e.g., the pyprocessing library
(included by defau
I'm taking the plunge and switching over to rpy2. While, Python to RObjects
seems straightforward, what about the reverse?
For example, I have a R function which returns a matrix. How can I get this
into a list of lists OR loop over each column (or row) pulling the vector
one at a time? What's
I have an R function that returns a list of Matrices and Vectors.
I've been processing the return list like:
rpyRV = rpy2.robjects.r.someFunc()
for property,rObj in zip([i for i in rpyRV.getnames()],[i for i in rpyRV]):
# do stuff with results
On my development box (64 bit, dual XEON) everythi
:59 PM, Mark Larsen wrote:
> I have an R function that returns a list of Matrices and Vectors.
>
> I've been processing the return list like:
>
> rpyRV = rpy2.robjects.r.someFunc()
> for property,rObj in zip([i for i in rpyRV.getnames()],[i for i in rpyRV]):
> # do
Hate to keep replying to my own message but...
I thought I'd throw a newer version of R at this problem. My above errors
are with 2.7.1 and I thought I'd test with 2.8.2 (my development box, where
all works well is at 2.7.2). I kept my main 2.7.1 install in /usr/lib64 and
installed 2.8.2 into /u
Thanks Laurent.
> Where was the hack needed ?
> Setting your target R as the first one found in the PATH should be
> enough.
You are correct, I didn't need to hack it, I wasn't sure. My hack was
to hardcode RHOMES.
A little progress. rpy2.robjects imports against R version 2.8.1 but
when I do
> ?! Weren't you on a Linux machine ?
> (aqua is for OS X, and your machine is amd64 - not something OS X is
> known to be officially working on)
Yes, Gentoo linux.
> Did you try having the install clearly distinct by using the option
> --prefix (and a setting PYTHONPATH accordingly) ?
I ended u
> Traceback (most recent call last):
> File "r_test.py", line 70, in ?
> r.barplot(bars, beside = "TRUE" )
> rpy.RPy_RException: Error in -0.01 * height : non-numeric argument to binary
> operator
Your Numeric array is being converted into a list.
>>> from rpy import r
>>> import Numeric
>>
> How do I get the matrix() function (or array for that matter) working
> properly for me!
Hmm, this works perfectly for me:
>>> import numpy
>>> from rpy import r
>>> import numpy
>>> bars = numpy.array([[25,15,26,25,18],[45,32,28,12,45]])
>>> r.barplot(bars)
[0.69996, 1.899
> table = r.table(wordsvector)
The table function returns a _contingency table_ object, which rpy
tries to unsuccessfully convert into a numpy array.
See if this returns something more workable:
rObj = with_mode(NO_CONVERSION, r.table)(words)
aHash = {}
[aHash.update(i) for i in rObj]
print aHas
> ImportError: No module named win32api
>
> What am I to do?
Install the win32 extensions:
http://python.net/crew/mhammond/win32/
--
Enter the BlackBerry Developer Challenge
This is your chance to win up to $100,000 in
> slope1 = with_mode(NO_CONVERSION, r.zyp.sen)(r("p~t"))
A quick glance tells me this line is a problem. The "." is reserved
in Python for modules/classes, try with an underscore:
slope1 = with_mode(NO_CONVERSION, r.zyp_sen)(r("p~t"))
> I don't understand what '*' and '**' mean. Could somebody let me know?
This is a python thing. '*' means that any number of arguments can be
passed as a tuple.
For instance, I have function:
def add(a,b):
return a+b
I could call this with a tuple argument:
var = (4,5)
add(*var)
'**' mean
> What is wrong?
Eva,
Make sure you are reading these docs for rpy2 here:
http://rpy.sourceforge.net/rpy2/doc/html/introduction.html#getting-started
And not the ones for rpy, they are very different implementations.
(IMHO, I'd stick with rpy2 over rpy).
You want this:
from rpy2 import robject
> If the above does not work, how do I access items in the returned list?
I frequently work like this:
an_r_list=robjects.r('list(one=1,two=2)')
list_names = [name for name in robjects.r.names(an_r_list)]
value, = an_r_list[list_names.index('one')]
---
> Hi, my problem is that when I use plot command, I can see for few second
> the graph but the window is closed quickly.
The python script is exiting and taking the R plot window with it.
Add, something like this to the end of your script:
raw_input("Press any Key To Exit")
>
> Apolagize if this is a stupid question. When I run the following script
>
> import rpy2.robjects as robjects
> x=[5.05,6.75,3.21,2.65]
> y=[1.65,2.64,2.64,6.95]
> print robjects.r.cor(x,y,method="spearman")
>
>
rpy2 is a lot more careful with automatic conversions. Try:
robjects.r.cor(rob
> model = rpy.r.glm(formula, data = d, family = 'binomial', control =
> rpy.r.glm.control(maxit = 500))
>
> Traceback (most recent call last):
> File "", line 1, in
> AttributeError: 'Robj' object has no attribute 'control'
It doesn't like the dot in the function name. Use:
model = rpy.r.glm(
19 matches
Mail list logo
