[Numpy-discussion] sparse array data

[Francesc Alted]

On 5/2/12 4:07 PM, Stéfan van der Walt wrote:
> Hi Francesc
>
> On Wed, May 2, 2012 at 1:53 PM, Francesc Alted<francesc at continuum.io>  wrote:
>> and add another one for the actual values of the array.  For a 3-D
>> sparse array, this looks like:
>>
>> dim0 | dim1 | dim2 | value
>> ==========================
>>     0 |   0  |   0  | val0
>>     0 |  10  | 100  | val1
>>    20 |   5  | 202  | val2
> What's the distinction between this and a coo_matrix?

Well, as the OP said, coo_matrix does not support dimensions larger than 
2, right?

In [4]: coo_matrix((3,4), dtype=np.int8).todense()
Out[4]:
matrix([[0, 0, 0, 0],
         [0, 0, 0, 0],
         [0, 0, 0, 0]], dtype=int8)


In [5]: coo_matrix((2,3,2), dtype=np.int8).todense()
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
/Users/faltet/<ipython-input-5-46b8ae62259f> in <module>()
----> 1 coo_matrix((2,3,2), dtype=np.int8).todense()

/Library/Frameworks/Python.framework/Versions/7.2/lib/python2.7/site-packages/scipy/sparse/coo.py 
in __init__(self, arg1, shape, dtype, copy)
     127                     obj, ij = arg1
     128                 except:
--> 129                     raise TypeError('invalid input format')
     130
     131                 try:

TypeError: invalid input format

-- 
Francesc Alted