josef.p...@gmail.com wrote:
> On Fri, Jun 5, 2009 at 1:48 AM, Robert Cimrman wrote:
>> josef.p...@gmail.com wrote:
>>> On Thu, Jun 4, 2009 at 4:30 PM, Gael Varoquaux
>>> wrote:
On Thu, Jun 04, 2009 at 10:27:11PM +0200, Kim Hansen wrote:
> "in(b)" or "in_iterable(b)" method, such that you
On Fri, Jun 5, 2009 at 1:48 AM, Robert Cimrman wrote:
> josef.p...@gmail.com wrote:
>> On Thu, Jun 4, 2009 at 4:30 PM, Gael Varoquaux
>> wrote:
>>> On Thu, Jun 04, 2009 at 10:27:11PM +0200, Kim Hansen wrote:
"in(b)" or "in_iterable(b)" method, such that you could do a.in(b)
which would
josef.p...@gmail.com wrote:
> On Thu, Jun 4, 2009 at 4:30 PM, Gael Varoquaux
> wrote:
>> On Thu, Jun 04, 2009 at 10:27:11PM +0200, Kim Hansen wrote:
>>> "in(b)" or "in_iterable(b)" method, such that you could do a.in(b)
>>> which would return a boolean array of the same shape as a with
>>> element
Anne Archibald wrote:
> 2009/6/4 :
>
>> intersect1d should throw a domain error if you give it arrays with
>> non-unique elements, which is not done for speed reasons
>
> It seems to me that this is the basic source of the problem. Perhaps
> this can be addressed? I realize maintaining compatibi
Kim Hansen wrote:
> Concerning the name setmember1d_nu, I personally find it quite verbose
> and not the name I would expect as a non-insider coming to numpy and
> not knowing all the names of the more special hidden-away functions
> and not being a python-wiz either.
To explain the naming: those
josef.p...@gmail.com wrote:
> On Thu, Jun 4, 2009 at 2:58 PM, Alan G Isaac wrote:
>> On 6/4/2009 1:27 PM josef.p...@gmail.com apparently wrote:
>>> Note: there are two versions of the docs for np.intersect1d, the
>>> currently published docs which describe the actual behavior (for the
>>> non-uniq
On Thu, Jun 4, 2009 at 4:52 PM, Gael Varoquaux
wrote:
> On Thu, Jun 04, 2009 at 04:43:39PM -0400, josef.p...@gmail.com wrote:
>> Just using "in" might promise more than it does, eg. it works only for
>> one dimensional arrays, maybe "in1d". With "in",
>
> Then 'in_1d'
No, if the breaks in a name
David Warde-Farley wrote:
> On 4-Jun-09, at 5:03 PM, Anne Archibald wrote:
>
>
>> Apart from the implementation issues people have chimed in about
>> already, it's worth noting that the speed of matrix multiplication
>> depends on the memory layout of the matrices. So generating B instead
>> dir
On 4-Jun-09, at 5:03 PM, Anne Archibald wrote:
> Apart from the implementation issues people have chimed in about
> already, it's worth noting that the speed of matrix multiplication
> depends on the memory layout of the matrices. So generating B instead
> directly as a 100 by 500 matrix might aff
Neal Becker wrote:
> Has this been considered as a candidate for our fft?
>
> http://sourceforge.net/projects/kissfft
>
I looked at it when I was looking for a BSD-compatible FFT with support
for prime factors (which fftpack does not handle). As Robert mentioned,
I did not see any compelling re
Matthew Brett wrote:
> Hi,
>
>
>>The RC1 for 0.7.1 scipy release has just been tagged. This is a
>> bug-only release
>>
>
> I feel (y)our pain, but don't you mean 'bug-fix only release'? ;-)
>
Actually, there is one big bug on python 2.6 for mac os x, so maybe the
bug-only is approp
On Jun 4, 2009, at 5:41 PM, Alan G Isaac wrote:
> On 6/4/2009 5:27 PM Tommy Grav apparently wrote:
>> Or the core development team split the matrices out of numpy and
>> make it
>> as separate package that the people that use them could pick up and
>> run with.
>
>
> This too would be a mistake,
On 6/4/2009 5:27 PM Tommy Grav apparently wrote:
> Or the core development team split the matrices out of numpy and make it
> as separate package that the people that use them could pick up and
> run with.
This too would be a mistake, I believe.
But it depends on whether a goal is to
have more
On Jun 4, 2009, at 5:25 PM, Christopher Barker wrote:
> Keith Goodman wrote:
>> Maybe announcing that numpy will drop support for matrices in a
>> future
>> version (3.0, ...) would save a lot of pain in the long run.
>
> Or make them better. There was a pretty good discussion of this a
> whil
Keith Goodman wrote:
> Maybe announcing that numpy will drop support for matrices in a future
> version (3.0, ...) would save a lot of pain in the long run.
Or make them better. There was a pretty good discussion of this a while
back on this list. We all had a lot of opinions, and there were some
2009/6/4 David Paul Reichert :
> Hi all,
>
> I would be glad if someone could help me with
> the following issue:
>
> From what I've read on the web it appears to me
> that numpy should be about as fast as matlab. However,
> when I do simple matrix multiplication, it consistently
> appears to be a
I should update after reading the thread Sebastian linked:
The current 1.3 version of numpy (don't know about previous versions) uses
the optimized Atlas BLAS routines for numpy.dot() if numpy was compiled with
these libraries. I've verified this on linux only, thought it shouldnt be
any different
Sebastian is right.
Since Matlab r2007 (i think that's the version) it has included support for
multi-core architecture. On my core2 Quad here at the office, r2008b has no
problem utilizing 100% cpu for large matrix multiplications.
If you download and build atlas and lapack from source and enab
On Thu, Jun 04, 2009 at 04:43:39PM -0400, josef.p...@gmail.com wrote:
> Just using "in" might promise more than it does, eg. it works only for
> one dimensional arrays, maybe "in1d". With "in",
Then 'in_1d'
> I found arraysetops because of unique1d, but I didn't figure out what
> the subpackage
On Thu, Jun 4, 2009 at 4:30 PM, Gael Varoquaux
wrote:
> On Thu, Jun 04, 2009 at 10:27:11PM +0200, Kim Hansen wrote:
>> "in(b)" or "in_iterable(b)" method, such that you could do a.in(b)
>> which would return a boolean array of the same shape as a with
>> elements true if the equivalent a members w
2009/6/4 :
> intersect1d should throw a domain error if you give it arrays with
> non-unique elements, which is not done for speed reasons
It seems to me that this is the basic source of the problem. Perhaps
this can be addressed? I realize maintaining compatibility with the
current behaviour is
On Thu, Jun 04, 2009 at 10:27:11PM +0200, Kim Hansen wrote:
> "in(b)" or "in_iterable(b)" method, such that you could do a.in(b)
> which would return a boolean array of the same shape as a with
> elements true if the equivalent a members were members in the iterable
> b.
That would really by what
Concerning the name setmember1d_nu, I personally find it quite verbose
and not the name I would expect as a non-insider coming to numpy and
not knowing all the names of the more special hidden-away functions
and not being a python-wiz either.
I think ain(a,b) would be the name I had expected as an
Thu, 04 Jun 2009 19:24:32 +0900, David Cournapeau wrote:
[clip]
> =
> SciPy 0.7.1 Release Notes
> =
>
> .. contents::
>
> SciPy 0.7.1 is a bug-fix release with no new features compared to 0.7.0.
scipy.special
=
Several bugs of varying
On Thu, Jun 04, 2009 at 11:03:36AM -0700, Ariel Rokem wrote:
>I have a question - how do I go about uninstalling my previous version of
>the ETS? A more general question to anyone - what's the right way of
>uninstalling any old python package? In the past, I have been advised to
>go
On Thu, Jun 4, 2009 at 2:58 PM, Alan G Isaac wrote:
> On 6/4/2009 1:27 PM josef.p...@gmail.com apparently wrote:
>> Note: there are two versions of the docs for np.intersect1d, the
>> currently published docs which describe the actual behavior (for the
>> non-unique case), and the new docs on the
On Jun 4, 2009, at 3:12 PM, Fernando Perez wrote:
> Howdy,
> I was thinking about this yesterday, because I'm dealing with exactly
> this same problem in a local project. How hard would it be to allow
> structured arrays to support ufuncs/arithmetic for the case where
> their dtype is actually
On Thu, Jun 4, 2009 at 13:30, Neal Becker wrote:
> Robert Kern wrote:
>
>> On Thu, Jun 4, 2009 at 11:58, Neal Becker wrote:
>>> Has this been considered as a candidate for our fft?
>>>
>>> http://sourceforge.net/projects/kissfft
>>
>> No. What would be the advantage of moving to Kiss FFT to offse
Howdy,
2009/6/3 Stéfan van der Walt :
>> however i seem to lose simple operations such as multiplication (a_array*2)
>> or powers (a_array**2).
>
> As a workaround, you can have two views on your data:
I was thinking about this yesterday, because I'm dealing with exactly
this same problem in a l
On 6/4/2009 1:27 PM josef.p...@gmail.com apparently wrote:
> Note: there are two versions of the docs for np.intersect1d, the
> currently published docs which describe the actual behavior (for the
> non-unique case), and the new docs on the doc editor
> http://docs.scipy.org/numpy/docs/numpy.lib.ar
Robert Kern wrote:
> On Thu, Jun 4, 2009 at 11:58, Neal Becker wrote:
>> Has this been considered as a candidate for our fft?
>>
>> http://sourceforge.net/projects/kissfft
>
> No. What would be the advantage of moving to Kiss FFT to offset the cost?
>
I was reading this:
http://listengine.tux
Hi -
thanks for all the good work on this!
I have been using an older version of the ETS, which I got when I installed
the EPD (4.0.30001) and now I have finally gotten around to trying to update
my ETS to this version.
I have a question - how do I go about uninstalling my previous version of
th
On Thu, Jun 4, 2009 at 11:58, Neal Becker wrote:
> Has this been considered as a candidate for our fft?
>
> http://sourceforge.net/projects/kissfft
No. What would be the advantage of moving to Kiss FFT to offset the cost?
--
Robert Kern
"I have come to believe that the whole world is an enigma
On Thu, Jun 4, 2009 at 12:32 PM, Alan G Isaac wrote:
> On 6/4/2009 11:29 AM josef.p...@gmail.com apparently wrote:
>> intersect1d is the intersection between sets (which are stored as
>> arrays), just like in the mathematical definition the two sets only
>> have unique elements
>
> Hmmm. OK, I se
Has this been considered as a candidate for our fft?
http://sourceforge.net/projects/kissfft
___
Numpy-discussion mailing list
Numpy-discussion@scipy.org
http://mail.scipy.org/mailman/listinfo/numpy-discussion
On 6/4/2009 12:08 PM Olivier Verdier apparently wrote:
> I really don't see any advantage of matrices over arrays for teaching. I
> prefer to teach linear algebra with arrays.
beta = (X.T*X).I * X.T * Y
beta = np.dot(np.dot(la.inv(np.dot(X.T,X)),X.T),Y)
I rest my case.
I would have to switch
On 6/4/2009 11:29 AM josef.p...@gmail.com apparently wrote:
> intersect1d is the intersection between sets (which are stored as
> arrays), just like in the mathematical definition the two sets only
> have unique elements
Hmmm. OK, I see you and Robert believe this.
But it does not match the do
I really don't see any advantage of matrices over arrays for teaching. I
prefer to teach linear algebra with arrays.
I would also like matrices to disappear from numpy. But then one would need
a new implementation of scipy.sparse, which is (very unfortunately)
matrix-based at the moment.
== Olivie
> On Sun, May 24, 2009 at 3:45 PM, David Warde-Farley
> wrote:
>> Anecdotally, it seems to me that lots of people (myself included) seem
>> to go through a phase early in their use of NumPy where they try to
>> use matrix(), but most seem to end up switching to using 2D arrays for
>> all the afor
On Sun, May 24, 2009 at 3:45 PM, David Warde-Farley wrote:
> Anecdotally, it seems to me that lots of people (myself included) seem
> to go through a phase early in their use of NumPy where they try to
> use matrix(), but most seem to end up switching to using 2D arrays for
> all the aforementione
On 4-Jun-09, at 9:28 AM, Stéfan van der Walt wrote:
> 2009/6/4 David Warde-Farley :
>> Sorry Pauli, seems I _don't_ have permission on the numpy trac to
>> change ticket status. The radio button shows up but then it gives
>> me a
>> "Warning: No permission to change ticket fields."
>
> Should be
2009/6/4 Matthew Brett :
>> The RC1 for 0.7.1 scipy release has just been tagged. This is a
>> bug-only release
>
> I feel (y)our pain, but don't you mean 'bug-fix only release'? ;-)
Thanks, guys! You made my weekend :-)
Cheers
Stéfan
___
Numpy-disc
After yesterdays discussion, I wanted to see if views of structured
arrays with mixed type can be easily used.
Is the following useful for the numpy user guide?
Josef
Calculations with mixed type structured arrays
--
>>> import numpy as np
The follo
On Thu, Jun 4, 2009 at 11:19 AM, Alan G Isaac wrote:
> On 6/4/2009 10:50 AM josef.p...@gmail.com apparently wrote:
>> intersect1d gives set intersection if both arrays have
>> only unique elements (i.e. are sets). I thought the
>> naming is pretty clear:
>
>> intersect1d(a,b) set intersection i
Alan G Isaac wrote:
> On 6/4/2009 10:50 AM josef.p...@gmail.com apparently wrote:
>> intersect1d gives set intersection if both arrays have
>> only unique elements (i.e. are sets). I thought the
>> naming is pretty clear:
>
>> intersect1d(a,b) set intersection if a and b with unique elements
On 6/4/2009 10:50 AM josef.p...@gmail.com apparently wrote:
> intersect1d gives set intersection if both arrays have
> only unique elements (i.e. are sets). I thought the
> naming is pretty clear:
> intersect1d(a,b) set intersection if a and b with unique elements
> intersect1d_nu(a,b) set
On Thu, Jun 4, 2009 at 11:12 AM, Alan G Isaac wrote:
>> On Thu, Jun 4, 2009 at 10:13 AM, Alan G Isaac wrote:
>>> Or if a stable order is not important (I don't
>>> recall if the OP specified), one could just
>>> np.intersect1d(a, np.unique(b))
>
> On 6/4/2009 10:50 AM josef.p...@gmail.com apparen
> On Thu, Jun 4, 2009 at 10:13 AM, Alan G Isaac wrote:
>> Or if a stable order is not important (I don't
>> recall if the OP specified), one could just
>> np.intersect1d(a, np.unique(b))
On 6/4/2009 10:50 AM josef.p...@gmail.com apparently wrote:
> This requires that also `a` has only unique elem
Hi,
> The RC1 for 0.7.1 scipy release has just been tagged. This is a
> bug-only release
I feel (y)our pain, but don't you mean 'bug-fix only release'? ;-)
Matthew
___
Numpy-discussion mailing list
Numpy-discussion@scipy.org
http://mail.scipy.org/ma
Thanks for the responses. I did not realize that dot() would do matrix
multiplication which was the main reason I was looking for a matrix-like
class. Like you and Tom suggested, I think it's best to stick to arrays.
Cheers,
Jason
On Sun, May 24, 2009 at 6:45 PM, David Warde-Farley wrote:
> O
On Thu, Jun 4, 2009 at 10:13 AM, Alan G Isaac wrote:
>> On Thu, Jun 4, 2009 at 8:23 AM, Alan G Isaac wrote:
>>> a[(a==b[:,None]).sum(axis=0,dtype=bool)]
>
>
> On 6/4/2009 8:35 AM josef.p...@gmail.com apparently wrote:
>> If b is large this creates a huge intermediate array
>
>
> True enough, but
On Thu, Jun 4, 2009 at 5:14 AM, David Cournapeau wrote:
> On Tue, Jun 2, 2009 at 10:56 PM, Ryan May wrote:
> > On Tue, Jun 2, 2009 at 5:59 AM, David Cournapeau
> > wrote:
> >>
> >> Robin wrote:
> >> > On Tue, Jun 2, 2009 at 11:36 AM, David Cournapeau >
> >> > wrote:
> >> >
> >> >> Done in r703
> On Thu, Jun 4, 2009 at 8:23 AM, Alan G Isaac wrote:
>> a[(a==b[:,None]).sum(axis=0,dtype=bool)]
On 6/4/2009 8:35 AM josef.p...@gmail.com apparently wrote:
> If b is large this creates a huge intermediate array
True enough, but one could then use fromiter:
setb = set(b)
itr = (ai for ai in a
Have a look at this thread:
http://www.mail-archive.com/numpy-discussion@scipy.org/msg13085.html
The speed difference is probably due to the fact that the matrix
multiplication does not call optimized an optimized blas routine, e.g.
the ATLAS blas.
Sebastian
On Thu, Jun 4, 2009 at 3:36 PM, Davi
Hi all,
I would be glad if someone could help me with
the following issue:
From what I've read on the web it appears to me
that numpy should be about as fast as matlab. However,
when I do simple matrix multiplication, it consistently
appears to be about 5 times slower. I tested this using
A = 0
2009/6/4 David Warde-Farley :
> Sorry Pauli, seems I _don't_ have permission on the numpy trac to
> change ticket status. The radio button shows up but then it gives me a
> "Warning: No permission to change ticket fields."
Should be fixed.
Cheers
Stéfan
___
On Thu, Jun 4, 2009 at 8:19 AM, wierob wrote:
> Hi,
>
> is there a BigInteger equivalent in numpy? The largest integer type I
> wound was dtype int64.
>
> I'm using stats.linregress to perform a regression analysis. The return
> stderr was nan because stas.ss(...) returned a negative number due to
On Thu, Jun 4, 2009 at 8:23 AM, Alan G Isaac wrote:
> a[(a==b[:,None]).sum(axis=0,dtype=bool)]
this is my preferred way when b is small and has unique elements.
if the elements in b are not unique, then be can be replaced by np.unique(b)
If b is large this creates a huge intermediate array
The a
a[(a==b[:,None]).sum(axis=0,dtype=bool)]
hth,
Alan Isaac
___
Numpy-discussion mailing list
Numpy-discussion@scipy.org
http://mail.scipy.org/mailman/listinfo/numpy-discussion
Hi,
is there a BigInteger equivalent in numpy? The largest integer type I
wound was dtype int64.
I'm using stats.linregress to perform a regression analysis. The return
stderr was nan because stas.ss(...) returned a negative number due to an
overflow. Setting dtype to int64 for my input data s
On 3-Jun-09, at 5:01 PM, Pauli Virtanen wrote:
>
> Btw, are you able to change the status of the ticket to
> "needs_review"?
> I think this should be possible for everyone, and not restricted to
> admins, but I'm not 100% sure...
Sorry Pauli, seems I _don't_ have permission on the numpy trac to
Hi,
The RC1 for 0.7.1 scipy release has just been tagged. This is a
bug-only release, see below for the release notes. More information can
also be found on the trac website:
http://projects.scipy.org/scipy/milestone/0.7.1
Please test it !
The scipy developers
--
=
On Tue, Jun 2, 2009 at 10:56 PM, Ryan May wrote:
> On Tue, Jun 2, 2009 at 5:59 AM, David Cournapeau
> wrote:
>>
>> Robin wrote:
>> > On Tue, Jun 2, 2009 at 11:36 AM, David Cournapeau
>> > wrote:
>> >
>> >> Done in r7031 - correlate/PyArray_Correlate should be unchanged, and
>> >> acorrelate/PyAr
Hi Neil,
Neil Crighton wrote:
> Hi all,
>
> I posted this message couple of days ago, but gmane grouped it with an old
> thread and it hasn't shown up on the front page. So here it is again...
>
> I'd really like to see the setmember1d_nu function in ticket 1036 get into
> numpy. There's a patc
64 matches
Mail list logo
