[Numpy-discussion] Re: numarray.objects missing in documentation

[Todd Miller]

Hi Faheem,

In order to create an object array, do this:

>>> import numarary.objects as obj
>>> a = obj.fromlist(["t","u","w",1])
>>> a
ObjectArray(["t","u","w",1)

Unfortunately, the ability to apply arbitrary functions to all the elements
of an object array has never been implemented.  Nevertheless, the following
might do:

def apply(array, function, result=None):
        rank1 = (array.rank == 1)
        if result is None:
                result0 = array.copy()
        else:
                result0 = result
        for x in range(array.shape[0]):
            if rank1:
                result0[x] = function(array[x])
            else:
                apply(array[x], function, result0[x])
        return result0

Re-ordering the loop and the if statement results in faster but slightly
more verbose code.

Regards,
Todd Miller


----- Original Message ----- 
From: "Faheem Mitha" <faheem at email.unc.edu>
To: <numpy-discussion at lists.sourceforge.net>
Sent: Sunday, March 14, 2004 12:20 PM
Subject: [Numpy-discussion] Re: numarray.objects missing in documentation


> On Sun, 14 Mar 2004 09:55:52 -0500, Todd Miller <jmiller at stsci.edu> wrote:
> > On Sun, 2004-03-14 at 00:26, Faheem Mitha wrote:
> >> Dear People,
> >>
> >> I'm posting this to gmane.comp.python.numeric.general. I am not sure
> >> what mailing list this corresponds to. I earlier tried to post this to
> >> the mailing list mentioned in the numarray documentation
> >> (numpy at lists.sourceforge.net), but it bounced.
> >>
> >> I don't see numarray.objects listed in
> >> http://stsdas.stsci.edu/numarray/numarray-0.9.html/genindex.html
> >
> > The section on object arrays has yet to be written.  What is it that you
> > want to know?
>
> Thanks for the quick reply. I'm trying to make an array of python
> objects as follows. Specifically, given a user defined class, I want
> to creata an array of instances of that class.
>
> Consider the following class.
>
> class cell:
>     def setrow(self,row):
>         self.row = row
>     def setcol(self,col):
>         self.col = col
>     ...
>
> I create an "empty" cell.
>
> empcell = cell()
>
> Now I want to create an array of objects identical to empcell, and
> later initialize them.
>
> I tried stuff along the lines of
>
> ******************************************************************
> import string
> from numarray import *
>
> empcell = cell()
>
> base = [empcell]
>
> mat = fromlist(base,shape=(2,2))
> ******************************************************************
>
> But the
>
> mat = fromlist(base,shape=(2,2))
>
> line fails with
>
> AttributeError: cell instance has no attribute '__len__'
>
> Similar things fail with the same error message. No doubt I'm doing
> something stupid, but would someone please enlighten me?
>
> I got the fromlist command from
>
> >>> help("numarray.objects")
>
> I was a bit disconcerted to find on trying to run the first example
> from the help page that I got
>
> >>> a = fromlist(["t","u","w"])
> ...
> TypeError: Expecting a python numeric type, got something else.
>
> What am I doing wrong?
>
> I'm running
>
> ii  python                    2.3.3-5                   An interactive
> high-level object-oriented language (default versio
> ii  python-numarray           0.8.1-1                   An array
> processing package modelled after Python-Numeric
> ii  python-numeric            23.1-1                    Numerical
> (matrix-oriented) Mathematics for Python
>
> on Debian sarge.
>
> Thanks in advance and sorry for the long-winded message.
>
>                                                           Faheem.
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> Numpy-discussion mailing list
> Numpy-discussion at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/numpy-discussion
>