[Numpy-discussion] Re: linalg.svd returns fortran arrays, causes problems

Travis Oliphant oliphant.travis at ieee.org
Mon Mar 20 19:51:05 CST 2006


Robert Kern wrote:
> I think this is a new bug, then.
>
> In [82]: reshape(numpy.array([[1,2,3,4,5,6,7,8,9,10],
> [11,12,13,14,15,16,17,18,19,20]], fortran = True), (2,5,2))
> Out[82]:
> array([[[ 1,  2],
>         [ 3,  4],
>         [ 5,  6],
>         [ 7,  8],
>         [ 9, 10]],
>
>        [[11, 12],
>         [13, 14],
>         [15, 16],
>         [17, 18],
>         [19, 20]]])
>
> In [83]: reshape(numpy.array([[1,2,3,4,5,6,7,8,9,10],
> [11,12,13,14,15,16,17,18,19,20]], fortran = False), (2,5,2))
> Out[83]:
> array([[[ 1,  2],
>         [ 3,  4],
>         [ 5,  6],
>         [ 7,  8],
>         [ 9, 10]],
>
>        [[11, 12],
>         [13, 14],
>         [15, 16],
>         [17, 18],
>         [19, 20]]])
>
>   
Robert,  can you test

numpy.array([[1,2,3,4,5,6,7,8,9,10],[11,12,13,14,15,16,17,18,19,20]], fortran = False).reshape(2,5,2)

numpy.array([[1,2,3,4,5,6,7,8,9,10],[11,12,13,14,15,16,17,18,19,20]], fortran = False).reshape(2,5,2)

please?

The reshape method has not changed since your version, but the reshape wrapper did change slightly, I believe.  In a manner that could have caused this difference. 

-Travis






More information about the Numpy-discussion mailing list