通过int选择numpy数组轴

Question

I am trying to access systematically a numpy array's axis. For example, suppose I have an array

a = np.random.random((10, 10, 10, 10, 10, 10, 10))
# choosing 7:9 from axis 2
b = a[:, :, 7:9, ...]
# choosing 7:9 from axis 3
c = a[:, :, :, 7:9, ...]

Typing colons gets very repetitive if I have a high dimensional array. Now, I want some function choose_from_axis such that

# choosing 7:9 from axis 2
b = choose_from_axis(a, 2, 7, 9)
# choosing 7:9 from axis 3
c = choose_from_axis(a, 3, 7, 9)

So, basically, I want to access an axis with a number. The only way I know how to do this is to use rollaxis back and forth, but I am looking for a more direct way to do it.

Answer 1

Sounds like you may be looking for take :

>>> a = np.random.randint(0,100, (3,4,5))
>>> a[:,1:3,:]
array([[[61,  4, 89, 24, 86],
        [48, 75,  4, 27, 65]],

       [[57, 55, 55,  6, 95],
        [19, 16,  4, 61, 42]],

       [[24, 89, 41, 74, 85],
        [27, 84, 23, 70, 29]]])
>>> a.take(np.arange(1,3), axis=1)
array([[[61,  4, 89, 24, 86],
        [48, 75,  4, 27, 65]],

       [[57, 55, 55,  6, 95],
        [19, 16,  4, 61, 42]],

       [[24, 89, 41, 74, 85],
        [27, 84, 23, 70, 29]]])

This will also give you support for tuple indexing. Example:

>>> a = np.arange(2*3*4).reshape(2,3,4)
>>> a
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]]])
>>> a[:,:,(0,1,3)]
array([[[ 0,  1,  3],
        [ 4,  5,  7],
        [ 8,  9, 11]],

       [[12, 13, 15],
        [16, 17, 19],
        [20, 21, 23]]])
>>> a.take((0,1,3), axis=2)
array([[[ 0,  1,  3],
        [ 4,  5,  7],
        [ 8,  9, 11]],

       [[12, 13, 15],
        [16, 17, 19],
        [20, 21, 23]]])

Answer 2

You could construct a slice object that does the job:

def choose_from_axis(a, axis, start, stop):
    s = [slice(None) for i in range(a.ndim)]
    s[axis] = slice(start, stop)
    return a[s]

For example, the following both give the same result:

x[:,1:2,:]
choose_from_axis(x, 1, 1, 2)

# [[[ 3  4  5]]
#  [[12 13 14]]
#  [[21 22 23]]]

as does the example in the question:

a = np.random.random((10, 10, 10, 10, 10, 10, 10))
a0 = a[:, :, 7:9, ...]
a1 = choose_from_axis(a, 2, 7, 9)

print np.all(a0==a1)   # True