Path: csiph.com!usenet.pasdenom.info!aioe.org!news.stack.nl!newsfeed.xs4all.nl!newsfeed1.news.xs4all.nl!xs4all!post.news.xs4all.nl!not-for-mail
To: python-list@python.org
From: Peter Otten <__peter__@web.de>
Subject: Re: Can anyone please help me in resolving the error => AttributeError: Array instance has no attribute '__trunc__'
Date: Tue, 09 Apr 2013 09:57:23 +0200
Organization: None
References: <c1cc2d05-e718-4414-a7f3-c1e03621c4f5@googlegroups.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7Bit
User-Agent: KNode/4.7.3
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.326.1365494162.3114.python-list@python.org>
Lines: 67
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: csiph.com comp.lang.python:43134

bhk755@gmail.com wrote:

> I am trying to create 2D arrays without using advanced features like
> numpy, 

I'd say that using ctypes is a bit more "advanced" than *using* numpy 
because with ctypes it helps to know C.

> for this I have created 2 separate modules arrays.py and
> array2D.py. Here's the code for that:
 
- You should absolutely go with numpy
- If you insist on building your own use lists of lists as suggested
- If you need something more space efficient than lists have a look at 

http://docs.python.org/2/library/array.html

rather than ctypes.

Now on to some debugging, without looking into the details.

> import arrays
> 
> class Array2D :
>     # Creates a 2-D array of size numRows x numCols.
>     def __init__( self, numRows, numCols ):
>     # Create a 1-D array to store an array reference for each row.
>                 
>         self._theRows = arrays.Array( numRows )

Here you create an instance of your array of integers...

>         # Create the 1-D arrays for each row of the 2-D array.
>         print "Num of Cloumns is", numCols
>                 
>         for i in range( numRows ) :
>             self._theRows[i] = arrays.Array( numCols )

...and here you are stuffing arrays.Array instances into that array of 
integers. In nuce:

>>> import ctypes
>>> class C: pass # classic class
... 
>>> items = (ctypes.c_int * 3)()
>>> items[0] = C()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: C instance has no attribute '__trunc__'

ctypes unhelpfully tries to truncate the value to convert it into an 
integer. As an aside here's an example of a successful conversion:

>>> class C2:
...     def __trunc__(self): return 42
... 
>>> items[0] = C2()
>>> items[0]
42

There are basically two resolutions:

- adjust the type of the outer array or
- use a single 1D array internally and calculate the index as
  row * width + column (for example).