On Tue, Jan 15, 2013 at 4:50 AM, Mads Ipsen <madsip...@gmail.com> wrote:
> Hi,
>
> I simply can't understand this. I'm trying to use argsort to produce
> indices that can be used to sort an array:
>
> from numpy import *
>
> indices = array([[4,3],[1,12],[23,7],[11,6],[8,9]])
> args = argsort(indices, axis=0)
> print indices[args]
>
> gives:
>
> [[[ 1 12]
> [ 4 3]]
>
> [[ 4 3]
> [11 6]]
>
> [[ 8 9]
> [23 7]]
>
> [[11 6]
> [ 8 9]]
>
> [[23 7]
> [ 1 12]]]
>
> I thought this should produce a sorted version of the indices array.
>
> Any help is appreciated.
>
>
Fancy indexing is a funny creature and not easy to understand in more than
one dimension. What is happening is that each index is replaced by the
corresponding row of a and the result is of shape (5,2,2). To do what you
want to do:
In [20]: a[i, [[0,1]]*5]
Out[20]:
array([[ 1, 3],
[ 4, 6],
[ 8, 7],
[11, 9],
[23, 12]])
I agree that there should be an easier way to do this.
Chuck
_______________________________________________
NumPy-Discussion mailing list
NumPy-Discussion@scipy.org
http://mail.scipy.org/mailman/listinfo/numpy-discussion
