[SciPy-user] Coo_constructor

[lorenzo bolla]

You can always define the sparse matrix dimension in this way:
scipy.sparse.coo_matrix(([5.], [[7.],[7.]]), dims=(16, 16))

L.

On Dec 14, 2007 8:04 PM, Nils Wagner <nwagner at iam.uni-stuttgart.de> wrote:

> Hi all,
>
> I use the coo constructor to build sparse matrices.
> A_i = coo_matrix( (vals_i, (rows_i, cols_i)) )
>
> I read the data for each matrix from a different file.
>
> One of the matrices is rank-one with only non-zero entry.
>
> (7, 7)        5.0
>
> Another matrix has the entries
>
>   (0, 0)        80000000.0
>   (0, 1)        -120000000.0
>   (1, 0)        -120000000.0
>   (1, 1)        480000000.0
>   (0, 2)        40000000.0
>   (2, 0)        40000000.0
>   (2, 2)        160000000.0
>   (1, 3)        -240000000.0
>   (3, 1)        -240000000.0
>   (2, 3)        -120000000.0
>   (3, 2)        -120000000.0
>   (3, 3)        480000000.0
>   (1, 4)        120000000.0
>   (4, 1)        120000000.0
>   (2, 4)        40000000.0
>   (4, 2)        40000000.0
>   (4, 4)        160000000.0
>   (3, 5)        -240000000.0
>   (5, 3)        -240000000.0
>   (4, 5)        -120000000.0
>   (5, 4)        -120000000.0
>   (5, 5)        480000000.0
>   (3, 6)        120000000.0
>   (6, 3)        120000000.0
>   (4, 6)        40000000.0
>   :     :
>   (9, 11)       -240000000.0
>   (11, 9)       -240000000.0
>   (10, 11)      -120000000.0
>   (11, 10)      -120000000.0
>   (11, 11)      480000000.0
>   (9, 12)       120000000.0
>   (12, 9)       120000000.0
>   (10, 12)      40000000.0
>   (12, 10)      40000000.0
>   (12, 12)      160000000.0
>   (11, 13)      -240000000.0
>   (13, 11)      -240000000.0
>   (12, 13)      -120000000.0
>   (13, 12)      -120000000.0
>   (13, 13)      480000000.0
>   (11, 14)      120000000.0
>   (14, 11)      120000000.0
>   (12, 14)      40000000.0
>   (14, 12)      40000000.0
>   (14, 14)      160000000.0
>   (13, 15)      120000000.0
>   (15, 13)      120000000.0
>   (14, 15)      40000000.0
>   (15, 14)      40000000.0
>   (15, 15)      80000000.0
>
>
> How can I circumvent the
> ValueError: shape mismatch: objects cannot be broadcast to
> a single shape
>
> if I try to add the matrices ?
>
> Nils
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.scipy.org/pipermail/scipy-user/attachments/20071214/45da190a/attachment.html>