Mailman 3 March 2011 - NumPy-Discussion

Partial least squares
by Nadav Horesh March 24, 2011

March 24, 2011

I am looking for a partial least sqaures code refactoring for two (X,Y) matrices. I found the following, but they not not work for me: 1. MDP: Factors only one matrix (am I wrong?) 2. pychem: Windows only code (I use Linux) 3. chemometrics from Koders: I get a singularity error. 4. pca_module (By Risvik): same problem as MDP Any suggestion? Nadav.

3 5

Question about repeated values in X in A[X] = Y
by Thouis (Ray) Jones March 24, 2011

March 24, 2011

I have done some work on scipy.ndimage.measurements, and was adding to it recently when I noticed I had made use of the idiom A[X] = Y with repeated indices in X, relying on the behavior that the last X,Y pair at each unique X is the one that is assigned in A (last = last by position in X). I thought I had found documentation of this behavior, somewhere, but when another developer asked, I was unable to locate it (perhaps I just forgot the google keywords I used last time). So, before I rely too much on this behavior, I thought I should make sure it was: - correct, - reasonable future-proof, - and documented (if the first two are true). Thanks, Thouis Jones

1 0

argmin and argmax without nan
by Dmitrey March 24, 2011

March 24, 2011

hi, is there any way to get argmin and argmax of an array w/o nans? Currently I have >>> from numpy import * >>> argmax([10,nan,100]) 1 >>> argmin([10,nan,100]) 1 But it's not the values I would like to get. The walkaround I use: get all indeces of nans, replace them by -inf, get argmax, replace them by inf, get argmin. Is there any better way? (BTW, I invoke argmin/argmax along of a chosen axis) D.

4 5

how to delete a particular column or row from numpy array based on some rule or value
by Sachin Kumar Sharma March 24, 2011

March 24, 2011

BB, In a numpy array of m x n size, I would like to delete few rows when a given element in that row is 'nan' or say any other value. For e.g. as in example given below, if I wish to delete a row when the 3rd element of row is zero, or if 3rd, 4th, 5th element are zero or either of 3rd, 4th and 5th element is zero. array([[ 1900. , nan, nan, nan, nan, nan], [ 1900.5, nan, nan, nan, nan, nan], [ 1901. , nan, nan, nan, nan, nan], ..., [ 6724. , nan, nan, nan, nan, nan], [ 6724.5, nan, nan, nan, nan, nan], [ 6725. , nan, nan, nan, nan, nan]]) Cheers Sachin ************************************************************************ Sachin Kumar Sharma Senior Geomodeler

3 2

Re: [Numpy-discussion] [NumPy-Tickets] [NumPy] #971: numpy.memmap 'offset' parameter docs are almost entirely wrong.
by Sebastian Haase March 24, 2011

March 24, 2011

Hi, I got an "Internal Error" message when trying to log into the Trac page. My comment: Could there at least be a note added to the current memmap doc string, stating that there is a problem with the current doc; also a link to http://projects.scipy.org/numpy/ticket/97 would be much better than nothing. My 2 cents, Sebastian Haase On Thu, Mar 24, 2011 at 1:58 AM, <numpy-tickets(a)scipy.org> wrote: > #971: numpy.memmap 'offset' parameter docs are almost entirely wrong. > ---------------------------+------------------------------------------------ > Reporter: 0ion9 | Owner: pv > Type: defect | Status: new > Priority: normal | Milestone: Unscheduled > Component: Documentation | Version: devel > Keywords: | > ---------------------------+------------------------------------------------ > Changes (by mwiebe): > > * milestone: => Unscheduled > > > -- > Ticket URL: <http://projects.scipy.org/numpy/ticket/971#comment:3> > NumPy <http://projects.scipy.org/numpy> > My example project > _______________________________________________ > NumPy-Tickets mailing list > NumPy-Tickets(a)scipy.org > http://mail.scipy.org/mailman/listinfo/numpy-tickets >

2 1

how to loop read input over various file with same extension in a folder (newbie question)
by Sachin Kumar Sharma March 24, 2011

March 24, 2011

BB, Currently I am reading a text file and storing the output as an array a1, some arithmetic operation on it and plotting it. I want to do the same over 1000 files. How to loop it and how to store the input in different arrays like a2, a3 and so on. Also advise what is the best way to TAG or assign a comment to an array to store which file it has read the data. Cheers Sachin ************************************************************************ Sachin Kumar Sharma Senior Geomodeler

2 2

Re: [Numpy-discussion] [SciPy-Dev] ANN: Numpy 1.6.0 beta 1
by Derek Homeier March 23, 2011

March 23, 2011

Hi Bruce, > Sorry as I should I have paid more attention to you first email as I > raised this on the numpy list. (At least with gmail, the reply is to > the > scipy-dev list rather than the numpy list.) > > Based on Pauli's answer I got around it by commenting out a line in > the > setup.py file about the git version: > http://mail.scipy.org/pipermail/numpy-discussion/2011-March/ > 055603.html > thanks, I was not subscribed to numpy-discussion (I realise I better should before going into major testing...)! This patch diff -ruN --exclude='*~' numpy-1.6.0b1.orig/setup.py numpy-1.6.0b1/ setup.py --- numpy-1.6.0b1.orig/setup.py 2011-03-23 09:24:00.000000000 +0100 +++ numpy-1.6.0b1/setup.py 2011-03-23 23:03:19.000000000 +0100 @@ -99,9 +99,10 @@ FULLVERSION = VERSION if os.path.exists('.git'): GIT_REVISION = git_version() -elif os.path.exists('numpy/version.py'): - # must be a source distribution, use existing version file - from numpy.version import git_revision as GIT_REVISION +# does not work in a clean source distribution: +#elif os.path.exists('numpy/version.py'): +# # must be a source distribution, use existing version file +# from numpy.version import git_revision as GIT_REVISION else: GIT_REVISION = "Unknown" did the build for Python3.1 and Python3.2. Cheers, Derek

1 0

Make array uncopyable
by Daniel Lepage March 23, 2011

March 23, 2011

Hi all, Is there a way to mark an array as uncopyable? If you have very large data arrays some numpy functions will cause MemoryErrors because behind the scenes they e.g. transpose the arrays into Fortran order to feed them into Fortran libraries (numpy.linalg.svd is an example of this). It would be great if there were a way to mark an array as "too big to copy" so that an attempt to call transpose() or astype() would raise an exception immediately instead of clobbering the machine's memory first, but I don't know of any flag that does this. I suppose I could always subclass ndarray and redefine transpose(), astype(), etc., but it'd be nice if there were an easier way. Thanks, Dan Lepage

2 3

np.histogramdd of empty data
by Nils Becker March 23, 2011

March 23, 2011

Hi, I was wondering why histogram2d and histogramdd raise a ValueError when fed with empty data of the correct dimensions. I came across this as a corner case when calling histogram2d from my own specialized histogram function. In comparison, histogram does handle this case correctly when bins are specified explicitely (obviously automatic binning cannot work without data...) np.histogram([], bins=([0,1])) (array([0]), array([0, 1])) np.histogram2d([],[], bins=([0,1],[0,1])) ValueError: zero-size array to ufunc.reduce without identity np.histogramdd([[],[]], bins=([0,1],[0,1])) ValueError: zero-size array to ufunc.reduce without identity cheers, nils

2 1

numpy log2 has bug
by Dmitrey March 23, 2011

March 23, 2011

>>> from numpy import log2, __version__ >>> log2(2**63) Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: log2 >>> __version__ '2.0.0.dev-1fe8136' (doesn't work with 1.3.0 as well)

3 2