[Numpy-discussion] SVD failure

[Stéfan van der Walt]

Hi Philipp

2009/5/5 Philipp K. Janert <python@beyondcode.org>:
> If I see this correctly, my SciPy version
> is 0.6.0; running on 64bit Suse 11.

SciPy 0.6 is quite old, and it is likely that the problem was fixed in
the mean time.

On SciPy 0.7 I see:

In [31]: u,s,v = linalg.svd(m)

In [32]: u
Out[32]:
array([[-0.70710678,  0.        , -0.70710678,  0.        ],
       [-0.70710678,  0.        ,  0.70710678,  0.        ],
       [ 0.        , -0.70710678,  0.        , -0.70710678],
       [ 0.        , -0.70710678,  0.        ,  0.70710678]])

In [33]: s
Out[33]: array([ 2.,  2.,  0.,  0.])

In [34]: v
Out[34]:
array([[-0.70710678, -0.70710678, -0.        , -0.        ],
       [-0.        , -0.        , -0.70710678, -0.70710678],
       [-0.70710678,  0.70710678,  0.        ,  0.        ],
       [ 0.        ,  0.        , -0.70710678,  0.70710678]])

Regards
Stéfan