I compiled the recent numpy from svn on gentoo-amd64, since then I often get
the error message:
RuntimeError: FATAL: module compiled aslittle endian, but detected different
endianness at runtime
Of course AMD64 is little endian, and array's byteorder is little endian by
default.
Grepping
On Tue, Sep 15, 2009 at 12:50:55AM -0400, Skipper Seabold wrote:
Doh, so it does! The docstring could probably note this unless I just
missed it somewhere.
Hey Skipper,
You sent a patch a while ago to fix a docstring. I am not sure it has
been applied ( :( ).
I just wanted to point out that
That it!
Thanks,
Nadav
-הודעה מקורית-
מאת: numpy-discussion-boun...@scipy.org בשם Citi, Luca
נשלח: ג 15-ספטמבר-09 11:32
אל: Discussion of Numerical Python
נושא: Re: [Numpy-discussion] Error in numpy 1.4.0 dev 07384
I got the same problem when compiling a new svn revision with
Hello,
I have a bit of code where I create arrays with meaningful names via:
meat = ['beef','lamb','pork']
cut = ['ribs','cutlets']
for m in meat:
for c in cut:
exec(consumed_%s_%s = np.zeros((numxgrid,numygrid,nummeasured)) %
(m,c))
Is this 'pythonic'? Or is it bad practice (and
hi John,
I have a bit of code where I create arrays with meaningful names via:
meat = ['beef','lamb','pork']
cut = ['ribs','cutlets']
for m in meat:
for c in cut:
exec(consumed_%s_%s = np.zeros((numxgrid,numygrid,nummeasured))
% (m,c))
Is this 'pythonic'? Or is it bad
On 9/15/2009 7:07 AM, Sebastien Binet wrote:
usage of the exec statement is usually frown upon and can be side stepped.
e.g:
for m in meat:
for c in cut:
locals()['consumed_%s_%s' % (m,c)] = some_array
Additionally, name construction can be pointless.
Maybe::
info = dict()
for
Hi Bruce
2009/9/14 Bruce Southey bsout...@gmail.com:
I would like to change the dtype of just one field of a structured or
record array without copying the original array. I can not change the
creation of the original array because it was created using genfromtxt.
You can't do that,
On 09/15/2009 07:46 AM, Stéfan van der Walt wrote:
Hi Bruce
2009/9/14 Bruce Southeybsout...@gmail.com:
I would like to change the dtype of just one field of a structured or
record array without copying the original array. I can not change the
creation of the original array because it
On Tue, Sep 15, 2009 at 3:37 AM, Gael Varoquaux
gael.varoqu...@normalesup.org wrote:
On Tue, Sep 15, 2009 at 12:50:55AM -0400, Skipper Seabold wrote:
Doh, so it does! The docstring could probably note this unless I just
missed it somewhere.
Hey Skipper,
You sent a patch a while ago to fix
On 09/14/2009 09:31 PM, Skipper Seabold wrote:
On Mon, Sep 14, 2009 at 9:59 PM, Pierre GMpgmdevl...@gmail.com wrote:
[snip]
OK, I see the problem...
When no dtype is defined, we try to guess what a converter should
return by testing its inputs. At first we check whether the input is a
Keep in mind that you can still have a problem with a conflict between your
SVN copy and system copy, if the SVN copy is visible by default (like, say,
installed to ~/.local under python 2.6) In my case, there was a problem
where a gnome panel applet used a feature in pygtk which called to numpy.
You're not supposed to write to the locals() dictionary. Sometimes
it works, but sometimes it doesn't. From the Python library docs:
locals()
Update and return a dictionary representing the current local symbol
table.
Note: The contents of this dictionary should not be
On Tue, Sep 15, 2009 at 9:43 AM, Bruce Southey bsout...@gmail.com wrote:
On 09/14/2009 09:31 PM, Skipper Seabold wrote:
On Mon, Sep 14, 2009 at 9:59 PM, Pierre GMpgmdevl...@gmail.com wrote:
[snip]
OK, I see the problem...
When no dtype is defined, we try to guess what a converter should
I spent much time on similar issues, so my policy became to keep all the
packages strictly under the same tree.
Nadav
-הודעה מקורית-
מאת: numpy-discussion-boun...@scipy.org בשם Ryan May
נשלח: ג 15-ספטמבר-09 17:08
אל: Discussion of Numerical Python
נושא: Re: [Numpy-discussion] Error
On Tue, Sep 15, 2009 at 10:44 AM, Skipper Seabold jsseab...@gmail.com wrote:
On Tue, Sep 15, 2009 at 9:43 AM, Bruce Southey bsout...@gmail.com wrote:
On 09/14/2009 09:31 PM, Skipper Seabold wrote:
On Mon, Sep 14, 2009 at 9:59 PM, Pierre GMpgmdevl...@gmail.com wrote:
[snip]
OK, I see the
On Tue, Sep 15, 2009 at 10:44 AM, Skipper Seabold jsseab...@gmail.com wrote:
On Tue, Sep 15, 2009 at 9:43 AM, Bruce Southey bsout...@gmail.com wrote:
On 09/14/2009 09:31 PM, Skipper Seabold wrote:
On Mon, Sep 14, 2009 at 9:59 PM, Pierre GMpgmdevl...@gmail.com wrote:
[snip]
OK, I see the
On Tuesday 15 September 2009 16:30:43 Rick White wrote:
You're not supposed to write to the locals() dictionary. Sometimes
it works, but sometimes it doesn't. From the Python library docs:
locals()
Update and return a dictionary representing the current local symbol
table.
On 09/15/2009 09:44 AM, Skipper Seabold wrote:
On Tue, Sep 15, 2009 at 9:43 AM, Bruce Southeybsout...@gmail.com wrote:
On 09/14/2009 09:31 PM, Skipper Seabold wrote:
On Mon, Sep 14, 2009 at 9:59 PM, Pierre GMpgmdevl...@gmail.comwrote:
[snip]
OK, I see the
Added: an inline grid
y,x = np.ogrid[ j:j+n, k:k+n ]
a[ j:j+n, k:k+n ] = f(x,y)
is 3* faster than a[y,x] = f(x,y) for 256x256,
about the same for little 8x8 squares (on mac ppc.)
So ogrids are not objects -- you can't
g = xxgrid[j:j+n, k:k+n]
...
use g, pass it just like
On Tue, Sep 15, 2009 at 09:22:41AM -0400, Skipper Seabold wrote:
Yes, of course. I have a login already, thanks. How quickly I
forget. I will have a look at the docs and add some examples.
Thanks a lot. Such contributions are very valuable to the community.
Gaël
hi,
when using numpy.random.multivariate_normal, would it make sense to warn
the user that they have entered a non-physical covariance matrix? i was
recently working on a problem and getting very strange results until i
finally realized that i had actually entered a bogus covariance matrix.
its
On Tue, Sep 15, 2009 at 11:38 AM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
hi,
when using numpy.random.multivariate_normal, would it make sense to warn
the user that they have entered a non-physical covariance matrix? i was
recently working on a problem and getting very strange
On Sep 15, 2009, at 10:44 AM, Skipper Seabold wrote:
How do you specify different dtypes in genfromtxt?
I could not see the information in the docstring and the dtype
argument
does not appear to allow multiple dtypes.
Just give a regular dtype, or something that could be interpreted as
On Tue, Sep 15, 2009 at 1:56 PM, Pierre GM pgmdevl...@gmail.com wrote:
On Sep 15, 2009, at 10:44 AM, Skipper Seabold wrote:
How do you specify different dtypes in genfromtxt?
I could not see the information in the docstring and the dtype
argument
does not appear to allow multiple dtypes.
On Tue, Sep 15, 2009 at 12:50, Charles R
Harrischarlesr.har...@gmail.com wrote:
On Tue, Sep 15, 2009 at 11:38 AM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
hi,
when using numpy.random.multivariate_normal, would it make sense to warn
the user that they have entered a non-physical
On Tue, 15 Sep 2009 13:26:23 -0500, Robert Kern wrote:
On Tue, Sep 15, 2009 at 12:50, Charles R
Harrischarlesr.har...@gmail.com wrote:
On Tue, Sep 15, 2009 at 11:38 AM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
hi,
when using numpy.random.multivariate_normal, would it
Hi there,
I have a user that failed to install numpy 1.3.0 on her Mac 10.5.8.
Turns out she is not using gfortran but g95.
Is it a known feature? Is g95 not supposed to work with numpy?
She did set FC to g95 before,
Here's the log.
Thanks,
C.
numpy.LOG
Description: Binary data
On Tue, Sep 15, 2009 at 12:57 PM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
On Tue, 15 Sep 2009 13:26:23 -0500, Robert Kern wrote:
On Tue, Sep 15, 2009 at 12:50, Charles R
Harrischarlesr.har...@gmail.com wrote:
On Tue, Sep 15, 2009 at 11:38 AM, Michael Gilbert
2009/9/15 Charles سمير Doutriaux doutria...@llnl.gov:
Hi there,
I have a user that failed to install numpy 1.3.0 on her Mac 10.5.8.
Turns out she is not using gfortran but g95.
Is it a known feature? Is g95 not supposed to work with numpy?
Probably not for Mac. Most likely no one has done
On Tue, 15 Sep 2009 13:17:43 -0600, Charles R Harris wrote:
On Tue, Sep 15, 2009 at 12:57 PM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
On Tue, 15 Sep 2009 13:26:23 -0500, Robert Kern wrote:
On Tue, Sep 15, 2009 at 12:50, Charles R
Harrischarlesr.har...@gmail.com wrote:
Thanks Robert,
That's exactly what I recommended her. Except I usually get gfortran
from http://hpc.sourceforge.net
C.
On Sep 15, 2009, at 12:24 PM, Robert Kern wrote:
2009/9/15 Charles سمير Doutriaux doutria...@llnl.gov:
Hi there,
I have a user that failed to install numpy 1.3.0 on her
On Tue, Sep 15, 2009 at 1:28 PM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
On Tue, 15 Sep 2009 13:17:43 -0600, Charles R Harris wrote:
On Tue, Sep 15, 2009 at 12:57 PM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
On Tue, 15 Sep 2009 13:26:23 -0500, Robert Kern wrote:
On Tue, 15 Sep 2009 13:31:21 -0600, Charles R Harris wrote:
On Tue, Sep 15, 2009 at 1:28 PM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
On Tue, 15 Sep 2009 13:17:43 -0600, Charles R Harris wrote:
On Tue, Sep 15, 2009 at 12:57 PM, Michael Gilbert
michael.s.gilb...@gmail.com
I just wanted to point out that there is an easy way of making a
difference, and making sure that the docstrings get fixed (which is
indeed very important). If you go to http://docs.scipy.org/ and register,
send your login name on this mailing list, we will add you to the list of
editors, and
Check the archives of the mailing list, there's an example using
dateutil.parser that may be just what you need.
How is this dateutil.parser used in timeseries?
Can it not be used to make the dateconverter obsolte for the most simple
cases?
___
On Tue, Sep 15, 2009 at 3:52 PM, Tim Michelsen
timmichel...@gmx-topmail.dewrote:
I just wanted to point out that there is an easy way of making a
difference, and making sure that the docstrings get fixed (which is
indeed very important). If you go to http://docs.scipy.org/ and
register,
On Sep 15, 2009, at 4:19 PM, Tim Michelsen wrote:
Check the archives of the mailing list, there's an example using
dateutil.parser that may be just what you need.
How is this dateutil.parser used in timeseries?
It's left in a corner. The use of dateutil.parser comes from
matplotlib.
I actually figured out a workaround with converters, since my missing
values are , ,ie., irregular number of spaces and the
values aren't stripped of white spaces. I just define {# : lambda s:
float(s.strip() or 0)}, and I have a loop build all of the converters,
but then I have to go
On Sep 15, 2009, at 4:22 PM, Ralf Gommers wrote:
There's a ticket for this functionality on the pydocweb tracker
already. Hopefully it gets implemented at some point.
My bad, sorry. I already always forget to check tickets on the trac
site for numpy/scipy, adding yet another site to check
On Sep 15, 2009, at 4:30 PM, Tim Michelsen wrote:
We shall all thank for having genfromtxt and derived!
You should really thank John Hunter, the original writer of
mlab.csv2rec (I thnk). I just polished the code and add a few extra
functionalities.
On Tue, Sep 15, 2009 at 4:33 PM, Pierre GM pgmdevl...@gmail.com wrote:
On Sep 15, 2009, at 4:22 PM, Ralf Gommers wrote:
There's a ticket for this functionality on the pydocweb tracker
already. Hopefully it gets implemented at some point.
My bad, sorry. I already always forget to check
It already exists in the docs:
http://docs.scipy.org/numpy/docs/numpy.lib.io.genfromtxt/
Did you mean you tried to edit this page?
But cannot be found in here:
http://docs.scipy.org/doc/numpy/reference/routines.io.html
So it just needs to be added in
On Tue, Sep 15, 2009 at 6:06 PM, Tim Michelsen
timmichel...@gmx-topmail.dewrote:
It already exists in the docs:
http://docs.scipy.org/numpy/docs/numpy.lib.io.genfromtxt/
Did you mean you tried to edit this page?
But cannot be found in here:
My bad, sorry. I already always forget to check tickets on the trac
site for numpy/scipy, adding yet another site to check seems to be way
too much for my caffeine-affected memory. However, there should be an
option to allocate tickets to specific people, right ? And then I
should
On Tue, Sep 15, 2009 at 12:26 PM, Robert Kern robert.k...@gmail.com wrote:
On Tue, Sep 15, 2009 at 12:50, Charles R
Harrischarlesr.har...@gmail.com wrote:
On Tue, Sep 15, 2009 at 11:38 AM, Michael Gilbert
michael.s.gilb...@gmail.com wrote:
hi,
when using
45 matches
Mail list logo