Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
Deschamps, Benjamin AGR.GC.CA> writes: > > R.dll is located at C:\Program Files\R\R-2.12.0\bin\i386\R.dll, > and my PATH includes C:\Program Files\R\R-2.12.0, C:\Program Files\R\R-2.12.0 \bin > and C:\Program Files\R\R-2.12.0\bin\i386. I also have the variable R_HOME > defined as C:\Program Files\R\R-2.12.0, and pywin32-214 is installed and working. > Typing “R” at the command prompt works, “R --version” works > and returns the correct version, and “R RHOME” returns C:\PROGA~1\R\R- 2.12~1.0 > (not sure why). By the way, I rpy2 was installed using the current windows > installer, rpy2-2.0.8.win32-py2.6.msi. Thanks, > > > > Benjamin > Has anyone solved this? I lost track of this thread, but now have the same problem. The recommended solution (changes to rinterface\__init__.py) does not work for me. The RPy2 error message I get refers to the proper R.dll path. The rinterface init module throws the error message anyway. The .dll allows all user full control so this should not be a file permission problem with "os.path.exists" -- Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev ___ rpy-list mailing list rpy-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rpy-list
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
On Tue, Nov 2, 2010 at 3:15 PM, John Schroeder wrote: > Deschamps, Benjamin AGR.GC.CA> writes: > > Has anyone solved this? I lost track of this thread, but now have the same > problem. The recommended solution (changes to rinterface\__init__.py) does > not work for me. The RPy2 error message I get refers to the proper R.dll > path. The rinterface init module throws the error message anyway. The .dll > allows all user full control so this should not be a file permission problem > with "os.path.exists" So you tried my suggestions to add os.path.join(R_HOME, 'bin', 'i386', 'R.dll') as a possible location for the DLL to rinterface\__init__.py Could you post the full traceback? Peter -- Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev ___ rpy-list mailing list rpy-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rpy-list
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
Yes,
Here is the trace back
Traceback (most recent call last):
File "E:\Users\roe\Eclipse\workspace\Python\RPy\src\Faithful.py", line
8, in
import rpy2.rpy_classic as rpy
File "C:\Python26\lib\site-packages\rpy2\rpy_classic.py", line 5, in
import rpy2.rinterface as ri
File "C:\Python26\lib\site-packages\rpy2\rinterface\__init__.py", line
56, in
raise RuntimeError("Unable to locate R.dll within %s" % R_HOME)
RuntimeError: Unable to locate R.dll within C:\Program
Files\R\R-2.12.0\bin\i386\
Here is the modified content of my __init__.py file. I am assuming I
delete the .pyc file to force a new one to be generated.
# Load the R dll using the explicit path
# First try the bin dir:
Rlib = os.path.join(R_HOME, 'bin', 'R.dll')
# Try bin/i386 subdirectory seen in R 2.12.0 ## ADDED ##
if not os.path.exists(Rlib): ## ADDED ##
Rlib = os.path.join(R_HOME, 'bin', 'i386', 'R.dll') ## ADDED ##
# Then the lib dir:
if not os.path.exists(Rlib):
Rlib = os.path.join(R_HOME, 'lib', 'R.dll')
#Rlib = os.path.join(R_HOME, 'R.dll')
# Otherwise fail out!
if not os.path.exists(Rlib):
raise RuntimeError("Unable to locate R.dll within %s" % R_HOME)
win32api.LoadLibrary( Rlib )
Here are key items from my environment
Path=C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\Program
Files\Co
mmon Files\Roxio
Shared\DLLShared\;C:\WINDOWS\system32\WindowsPowerShell\v1.0;C:
\Documents and Settings\roe\Application
Data\Python\Scripts;C:\Python26\Scripts;
C:\Python26\;C:\MinGW\bin;C:\MinGW\msys\bin;C:\Program
Files\R\R-2.12.0\bin\i386
R_HOME=C:\Program Files\R\R-2.12.0\bin\i386\
Here is the location of my R.dll and .exe
C:\Program Files\R\R-2.12.0\bin\i386
John A. Schroeder
Idaho National Laboratory
Battelle Energy Alliance, LLC
P.O. Box 1625
Idaho Falls, ID 83415-3850
Ph: 208-526-8755
FAX: 208-526-2930
Peter
11/02/2010 09:48 AM
Please respond to
rpy-list@lists.sourceforge.net
To
"RPy help, support and design discussion list"
cc
Subject
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
On Tue, Nov 2, 2010 at 3:15 PM, John Schroeder
wrote:
> Deschamps, Benjamin AGR.GC.CA> writes:
>
> Has anyone solved this? I lost track of this thread, but now have the
same
> problem. The recommended solution (changes to rinterface\__init__.py)
does
> not work for me. The RPy2 error message I get refers to the proper
R.dll
> path. The rinterface init module throws the error message anyway. The
.dll
> allows all user full control so this should not be a file permission
problem
> with "os.path.exists"
So you tried my suggestions to add os.path.join(R_HOME, 'bin', 'i386',
'R.dll')
as a possible location for the DLL to rinterface\__init__.py
Could you post the full traceback?
Peter
--
Nokia and AT&T present the 2010 Calling All Innovators-North America
contest
Create new apps & games for the Nokia N8 for consumers in U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in
marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
http://p.sf.net/sfu/nokia-dev2dev
___
rpy-list mailing list
rpy-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rpy-list
--
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
http://p.sf.net/sfu/nokia-dev2dev___
rpy-list mailing list
rpy-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rpy-list
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
OOPs,
Looks like it is working if I correct my R_HOME from
R_HOME=C:\Program Files\R\R-2.12.0\bin\i386\
to
R_HOME=C:\Program Files\R\R-2.12.0
The incorrect R_HOME was a result of an earlier attempt to get this
working. Your fix appears to do the job
Thankyou!
And sorry about the interuption due to a goof on my part!
John A. Schroeder
Idaho National Laboratory
Battelle Energy Alliance, LLC
P.O. Box 1625
Idaho Falls, ID 83415-3850
Ph: 208-526-8755
FAX: 208-526-2930
John A Schroeder
11/02/2010 10:04 AM
Please respond to
"RPy help, support and design discussion list"
To
rpy-list@lists.sourceforge.net
cc
Subject
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
Yes,
Here is the trace back
Traceback (most recent call last):
File "E:\Users\roe\Eclipse\workspace\Python\RPy\src\Faithful.py", line
8, in
import rpy2.rpy_classic as rpy
File "C:\Python26\lib\site-packages\rpy2\rpy_classic.py", line 5, in
import rpy2.rinterface as ri
File "C:\Python26\lib\site-packages\rpy2\rinterface\__init__.py", line
56, in
raise RuntimeError("Unable to locate R.dll within %s" % R_HOME)
RuntimeError: Unable to locate R.dll within C:\Program
Files\R\R-2.12.0\bin\i386\
Here is the modified content of my __init__.py file. I am assuming I
delete the .pyc file to force a new one to be generated.
# Load the R dll using the explicit path
# First try the bin dir:
Rlib = os.path.join(R_HOME, 'bin', 'R.dll')
# Try bin/i386 subdirectory seen in R 2.12.0 ## ADDED ##
if not os.path.exists(Rlib): ## ADDED ##
Rlib = os.path.join(R_HOME, 'bin', 'i386', 'R.dll') ## ADDED ##
# Then the lib dir:
if not os.path.exists(Rlib):
Rlib = os.path.join(R_HOME, 'lib', 'R.dll')
#Rlib = os.path.join(R_HOME, 'R.dll')
# Otherwise fail out!
if not os.path.exists(Rlib):
raise RuntimeError("Unable to locate R.dll within %s" % R_HOME)
win32api.LoadLibrary( Rlib )
Here are key items from my environment
Path=C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\Program
Files\Co
mmon Files\Roxio
Shared\DLLShared\;C:\WINDOWS\system32\WindowsPowerShell\v1.0;C:
\Documents and Settings\roe\Application
Data\Python\Scripts;C:\Python26\Scripts;
C:\Python26\;C:\MinGW\bin;C:\MinGW\msys\bin;C:\Program
Files\R\R-2.12.0\bin\i386
R_HOME=C:\Program Files\R\R-2.12.0\bin\i386\
Here is the location of my R.dll and .exe
C:\Program Files\R\R-2.12.0\bin\i386
John A. Schroeder
Idaho National Laboratory
Battelle Energy Alliance, LLC
P.O. Box 1625
Idaho Falls, ID 83415-3850
Ph: 208-526-8755
FAX: 208-526-2930
Peter
11/02/2010 09:48 AM
Please respond to
rpy-list@lists.sourceforge.net
To
"RPy help,support and design discussion list"
cc
Subject
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
On Tue, Nov 2, 2010 at 3:15 PM, John Schroeder
wrote:
> Deschamps, Benjamin AGR.GC.CA> writes:
>
> Has anyone solved this? I lost track of this thread, but now have the
same
> problem. The recommended solution (changes to rinterface\__init__.py)
does
> not work for me. The RPy2 error message I get refers to the proper
R.dll
> path. The rinterface init module throws the error message anyway. The
.dll
> allows all user full control so this should not be a file permission
problem
> with "os.path.exists"
So you tried my suggestions to add os.path.join(R_HOME, 'bin', 'i386',
'R.dll')
as a possible location for the DLL to rinterface\__init__.py
Could you post the full traceback?
Peter
--
Nokia and AT&T present the 2010 Calling All Innovators-North America
contest
Create new apps & games for the Nokia N8 for consumers in U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in
marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
http://p.sf.net/sfu/nokia-dev2dev
___
rpy-list mailing list
rpy-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rpy-list
--
Nokia and AT&T present the 2010 Calling All Innovators-North America
contest
Create new apps & games for the Nokia N8 for consumers in U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in
marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
http://p.sf.net/sfu/nokia-dev2dev
___
rpy-list mailing list
rpy-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rpy-list
--
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in U.S. an
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
On Tue, Nov 2, 2010 at 4:11 PM, John A Schroeder wrote: > > OOPs, > > > Looks like it is working if I correct my R_HOME from > > R_HOME=C:\Program Files\R\R-2.12.0\bin\i386\ > > to > > R_HOME=C:\Program Files\R\R-2.12.0 > > The incorrect R_HOME was a result of an earlier attempt to get this working. > Your fix appears to do the job > > Thankyou! Well spotted - I was just writing a reply to clarify if you'd set R_HOME like that. I'm glad the fix works - Laurent, can you apply this to the repository? Peter -- Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev ___ rpy-list mailing list rpy-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rpy-list
Re: [Rpy] rpy2 2.0.8 win32 Unable to locate R.dll
On 02/11/10 17:22, Peter wrote: > On Tue, Nov 2, 2010 at 4:11 PM, John A Schroeder > wrote: >> OOPs, >> >> >> Looks like it is working if I correct my R_HOME from >> >> R_HOME=C:\Program Files\R\R-2.12.0\bin\i386\ >> >> to >> >> R_HOME=C:\Program Files\R\R-2.12.0 >> >> The incorrect R_HOME was a result of an earlier attempt to get this working. >> Your fix appears to do the job >> >> Thankyou! > Well spotted - I was just writing a reply to clarify if you'd set R_HOME like > that. I'm glad the fix works - Laurent, can you apply this to the repository? > Thanks for the patch Peter. I just pushed to the branch "version_2.0.x", entry 1326:4af6fc146f3. Someone will have to contribute the binary builds. I have not bothered to apply to other branches. Laurent > Peter > > -- > Nokia and AT&T present the 2010 Calling All Innovators-North America contest > Create new apps& games for the Nokia N8 for consumers in U.S. and Canada > $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing > Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store > http://p.sf.net/sfu/nokia-dev2dev > ___ > rpy-list mailing list > rpy-list@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/rpy-list -- Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev ___ rpy-list mailing list rpy-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rpy-list
Re: [Rpy] converting r array to numpy array
Thanks Laurent,
that makes sense. I noticed someone reported this problem earlier (I
probably should have looked there first!), but didn't follow it up so
their issue got closed. I reopened it, hope that's ok. See issue #47
rpy2 -> numpy mangles arrays.
Thanks again,
Robert
-Original Message-
From: Laurent Gautier [mailto:lgaut...@gmail.com]
Sent: Tuesday, 2 November 2010 4:49 PM
To: Denham Robert
Cc: RPy help, support and design discussion list
Subject: Re: [Rpy] converting r array to numpy array
This is a bug then.
The problem is in the way strides are computed (buffer.c in
rpy/rinterface), and a fix seems to be:
static void
sexp_strides(SEXP sexp, Py_intptr_t *strides, Py_ssize_t itemsize,
Py_intptr_t *shape, int nd)
{
/* Set the buffer 'strides', that is a vector or Py_intptr_t
* containing the offset (in bytes) when progressing along
* each dimension.
*/
int i;
Py_intptr_t cumul_strides = 1;
strides[0] = itemsize;
for (i = 1; i < nd; i++) {
strides[i] = shape[i-1] * strides[i-1];
}
}
I just committed that change to the branch 2.2.x. but I'd like to have
this tested a little before porting it back to 2.1.x.
Can you fill one on the bitbucket tracker ? (so others know about the
issue)
Thanks,
L.
On 01/11/10 22:36, Denham Robert wrote:
> Yeah, sorry, this is the output:
> In [3]: import numpy
> In [4]: import rpy2
> In [5]: import rpy2.robjects as robjects In [7]: marr =
> robjects.r.array(robjects.IntVector(range(54)),dim=
> robjects.IntVector([9,2,3]))
>
> In [8]: marrpy = numpy.array(marr)
> In [9]: print marrpy.flatten()
> --> print(marrpy.flatten())
> [ 0 2 4 9 11 13 1 3 5 10 12 14 2 4 6 11 13 15 3 5 7 12 14
> 16
> 4
>6 8 13 15 17 5 7 9 14 16 18 6 8 10 15 17 19 7 9 11 16 18 20
> 8 10
> 12 17 19 21]
>
> In [10]: print robjects.r['as.vector'](marr)
> ---> print(robjects.r['as.vector'](marr))
> [1] 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
> 21
> 22 23 24
> [26] 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
> 47 48 49
> [51] 50 51 52 53
>
>
> In [15]: print marrpy.flatten('F')
> ---> print(marrpy.flatten('F'))
> [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 2 3 4 5 6
> 7
> 8
>9 10 11 12 13 14 15 16 17 18 19 4 5 6 7 8 9 10 11 12 13 14 15
> 16
> 17
> 18 19 20 21]
>
>
> I was prepared for differences due to column/row ordering, but this
> doesn't seem to be the problem.
>
> Works fine on matrices:
>
> In [11]: mat = robjects.r.matrix(robjects.IntVector(range(54)),nrow=9)
> In [12]: matpy = numpy.array(mat)
> In [13]: print mat
> ---> print(mat)
>[,1] [,2] [,3] [,4] [,5] [,6]
> [1,]09 18 27 36 45
> [2,]1 10 19 28 37 46
> [3,]2 11 20 29 38 47
> [4,]3 12 21 30 39 48
> [5,]4 13 22 31 40 49
> [6,]5 14 23 32 41 50
> [7,]6 15 24 33 42 51
> [8,]7 16 25 34 43 52
> [9,]8 17 26 35 44 53
>
>
> In [14]: print matpy
> ---> print(matpy)
> [[ 0 9 18 27 36 45]
> [ 1 10 19 28 37 46]
> [ 2 11 20 29 38 47]
> [ 3 12 21 30 39 48]
> [ 4 13 22 31 40 49]
> [ 5 14 23 32 41 50]
> [ 6 15 24 33 42 51]
> [ 7 16 25 34 43 52]
> [ 8 17 26 35 44 53]]
>
>
>
> Regards, Robert
>
>
>
> -Original Message-
> From: Laurent Gautier [mailto:lgaut...@gmail.com]
> Sent: Monday, 1 November 2010 4:58 PM
> To: Denham Robert
> Cc: RPy help, support and design discussion list
> Subject: Re: [Rpy] converting r array to numpy array
>
> On 01/11/10 07:28, Denham Robert wrote:
>> I got caught out converting a 3d array in R to a numpy array. I did
>> something like:
>>
>> import numpy
>> import rpy2
>> import rpy2.robjects as robjects
>>
>> marr = robjects.r.array(robjects.IntVector(range(54)),dim=
>> robjects.IntVector([9,2,3]))
>> marrpy = numpy.array(marr)
>> print marrpy.flatten()
>> print robjects.r['as.vector'](marr)
>>
>> Which clearly doesn't work (at least for me, using rpy2.1.7, R-10.0
>> on 64bit suse linux).
> Could you be more specific about "does not work" ?
> If you are after getting the same output for the two print calls, this
> is not happening because arrays are being stored "column-major" by R
> by while numpy has "row-major" as a default.
> Try:
>
> print marrpy.flatten('F')
>
>
>
>>The matrix version does work, eg
>>
>> mat = robjects.r.matrix(robjects.IntVector(range(54)),nrow=9)
>> matpy = numpy.array(mat)
>> print mat
>> print matpy
>>
>> I checked the documentation, and found examples for vectors and
>> matrices, but not for arrays, but also no warnings. So, I was just
>> wondering, is this supposed to work? If it is, a hint on why it
>> doesn't work for me would be good. If it isn't supposed to work, I
>> was wondering if it were possible to throw an error or something.
>
> It is supposed to work, but not claim about an absence of bugs is
made.
>
>
> Bes
