On 3/15/11 8:28 PM, Brent Pedersen wrote:
> system info below.
> maybe it's internal to R:
>
>
>> a = 1:199999999
>> sum(a)
> [1] NA
> Warning message:
> In sum(a) : Integer overflow - use sum(as.numeric(.))
>> sum(as.numeric(a))
> Error: cannot allocate vector of size 1.5 Gb

This is a limitation/characteristic of R, somehow. Integers are C-int 
while in Python they are C-long in Python >= 3, C-int/C-long in Python < 
3 (I am not sure about where Python 2.7 stands).
Where Python < 3 would silently upgrade an int to a long, R's sum throws 
an overflow error.

>
>
> but this works in a new session:
>
>>   a = vector(length=199999999, mode="numeric")
>> sum(a)
> [1] 0

You are requiring less RAM because you do not have to hold two long 
vectors in memory at the same time. as.numeric() does make a copy.

>
>
> ###############################################
>
>

import numpy
numpy.__version__ ?

> Python 2.6.6 (r266:84292, Sep 15 2010, 15:52:39)
> [GCC 4.4.5] on linux2
>
> R version 2.12.1 (2010-12-16)
> Copyright (C) 2010 The R Foundation for Statistical Computing
> ISBN 3-900051-07-0
>
> Linux latitude 2.6.35-27-generic-pae #48-Ubuntu SMP Tue Feb 22
> 21:46:58 UTC 2011 i686 GNU/Linux
>
> $ hg branch; hg identify
> version_2.2.x
> 97d788ba27f8+ (version_2.2.x) tip

Good. Eventual fixes would appear there first.


>
> On Tue, Mar 15, 2011 at 1:03 PM, Laurent Gautier<lgaut...@gmail.com>  wrote:
>> On 3/15/11 5:45 PM, Brent Pedersen wrote:
>>>
>>> On Tue, Mar 15, 2011 at 9:40 AM, Laurent Gautier<lgaut...@gmail.com>
>>>   wrote:
>>>>
>>>> It can do either one of the two.
>>>>
>>>>
>>>> http://rpy.sourceforge.net/rpy2/doc-dev/html/numpy.html?#low-level-interface
>>>>
>>>>
>>>>
>>>> L.
>>>>
>>>
>>> how about in this case?::
>>>
>>>
>>>    import numpy as np
>>>    import rpy2.robjects as robjects
>>>    import rpy2.robjects.numpy2ri
>>>    R = robjects.r
>>>
>>>    a = np.arange(199999999)
>>>    R.sum(a)
>>>
>>> it seems to always make a copy, even if i have it in column order.
>>
>> It shouldn't be the case, I think. Any details about the Python/numpy/rpy2
>> versions ?
>>
>>


------------------------------------------------------------------------------
Colocation vs. Managed Hosting
A question and answer guide to determining the best fit
for your organization - today and in the future.
http://p.sf.net/sfu/internap-sfd2d
_______________________________________________
rpy-list mailing list
rpy-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rpy-list

Reply via email to