Please keep the thread on the list
<I'll do so as this does not appear to contain anything that shouldn't>
On 1/13/10 5:29 PM, Nick Schurch wrote:
> 2010/1/13 Laurent Gautier<lgaut...@gmail.com>:
>>
>>
>>
>> Why use "True" and "False" when Python has True and False ?
>>
>> Why define the function and assign it to an R variable name at each call ?
>
> This was an example only - the actual subroutine is rather complex and
> so is the r function that is built. I have found that for complex sets
> of R commands, assigning them to a function is the most reliable way
> of being able to deal things and, of course, I want to be able to use
> the function repeatedly within a subroutine (ideally).
Combining R commands into an R function is fine, but that something that
can be distinct from assigning a R object (here a function) to a
symbol/name (here "filter <- ").
For example in rpy2 (in rpy that the same with the NO_CONVERSION mode I
think):
>>> import rpy2.robjects as robjects
>>> f = robjects.r("function(x) x^2")
>>> tuple(f(4))
(16,)
The function has for "Python name" "f", and is exposed a Python
function-like object, but does not have an "R name".
Now consider the following Python function:
def g(x, want_f=True):
if want_f:
return f(x)
else:
None
>> There is the remote possibility that there is caching involved, and r.filter
>> is bound to your first function definition (and not rebound when
>> redefined... since rpy is not told to do so).
>> Anyway, what is shown can be seen as not-so-elegant; what about defining 2
>> functions filter_true, filter_false one for all call one or the other. Or
>> even better, make named list element ("forenames" or "surnames" a parameter
>> to *one* function) ?
>
> As I said, the actual subroutine is rather complex and currently does
> define multiple functions for the different cases - but I'm up to 6
> different functions now and it really isn't an elegant solution. There
> must be a way to force rpy to unbind a function.
An elegant way would not involve overwriting the same global R variable
over and over again, IMHO... especially if you have complex routines
calling eachothers (and possibly overwriting each other's previously
defined R functions).
>> In R, one would delete the object going by the symbol "filter" from
>> GlobalEnv:
>>
>> rm("filter")
>
> This doesn't seem to work from inside python though.
>
I might be mistaken, but I do not see why it wouldn't. How are you
trying to achieve it (a standalone example one could copy/paste would
demonstrate that it is not working).
L.
------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
rpy-list mailing list
rpy-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rpy-list
