On Tue, Dec 13, 2011 at 4:11 PM, Ken Basye <kbas...@jhu.edu> wrote:
> Hi folks,
> I need an efficient way to get both the min and argmin of a 2-d
> array along one axis. It seemed to me that the way to do this was to
> get the argmin and then use it to index into the array to get the min,
> but I can't figure out how to do it. Here's my toy example:
>
> >>> x = np.arange(25).reshape((5,5))
> >>> x
> array([[ 0, 1, 2, 3, 4],
> [ 5, 6, 7, 8, 9],
> [10, 11, 12, 13, 14],
> [15, 16, 17, 18, 19],
> [20, 21, 22, 23, 24]])
> >>> y = np.abs(x - x.T)
> >>> y
> array([[ 0, 4, 8, 12, 16],
> [ 4, 0, 4, 8, 12],
> [ 8, 4, 0, 4, 8],
> [12, 8, 4, 0, 4],
> [16, 12, 8, 4, 0]])
> >>> np.argmin(y, axis=0)
> array([0, 1, 2, 3, 4])
> >>> np.min(y, axis=0)
> array([0, 0, 0, 0, 0])
>
> Here it seems like there should be a simple way to get the same array
> that min() returns using the argmin result, which won't need to 'search'
> in the array.
>
>
You can use the result of argmin to index into y, if you combine it with,
say, arange(ncols) in the second dimension:
In [53]: y = random.randint(0,10,size=(5,7))
In [54]: y
Out[54]:
array([[3, 3, 5, 1, 5, 3, 7],
[1, 0, 6, 8, 0, 1, 1],
[7, 9, 9, 3, 3, 1, 6],
[5, 3, 5, 4, 9, 7, 4],
[1, 7, 1, 6, 6, 1, 8]])
In [55]: am = np.argmin(y, axis=0)
In [56]: am
Out[56]: array([1, 1, 4, 0, 1, 1, 1])
In [57]: colmins = y[am, arange(7)]
In [58]: colmins
Out[58]: array([1, 0, 1, 1, 0, 1, 1])
Warren
_______________________________________________
NumPy-Discussion mailing list
NumPy-Discussion@scipy.org
http://mail.scipy.org/mailman/listinfo/numpy-discussion
