理解numpy ndarry中的axis

理解shape()中对应的axis

arr = np.arange(0,24).reshape(2,3,4)
arr

打印：

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]]])

各个axis上的元素是什么

• shape中的2，3，4分别对应axis中的0，1，2
• axis0有2个元素，分别是

arr[0,:]
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])

arr[1,:]
array([[12, 13, 14, 15],
       [16, 17, 18, 19],
       [20, 21, 22, 23]])

• axis1中有3个元素

arr[:,0]
array([[ 0,  1,  2,  3],
       [12, 13, 14, 15]])

arr[:,1]
array([[ 4,  5,  6,  7],
       [16, 17, 18, 19]])

arr[:,2]
array([[ 8,  9, 10, 11],
       [20, 21, 22, 23]])

• axis2中有4个元素

arr[:,:,0]
array([[ 0,  4,  8],
       [12, 16, 20]])

arr[:,:,1]
array([[ 1,  5,  9],
       [13, 17, 21]])

arr[:,:,2]
array([[ 2,  6, 10],
       [14, 18, 22]])

arr[:,:,3]
array([[ 3,  7, 11],
       [15, 19, 23]])

对axis的操作，np.delete(ndarray, index, axis)

arr0 = np.delete(arr, 0, axis=0)
arr0
array([[[12, 13, 14, 15],
        [16, 17, 18, 19],
        [20, 21, 22, 23]]])

arr1 = np.delete(arr, 0, axis=1)
arr1
array([[[ 4,  5,  6,  7],
        [ 8,  9, 10, 11]],

       [[16, 17, 18, 19],
        [20, 21, 22, 23]]])

arr2 = np.delete(arr, 0, axis=2)
arr2
array([[[ 1,  2,  3],
        [ 5,  6,  7],
        [ 9, 10, 11]],

       [[13, 14, 15],
        [17, 18, 19],
        [21, 22, 23]]])

一个方便记忆axis的方法，shape(2,3,4)中的各个数字对应数组从外到里面的方括号数。看例子：

arr
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]]])

• axis=0，从最外层进入一个方括号，就有两对[[ … ]]
• axis=1,从外层进入进两个方括号，有三对[…]
• axis=2,就是数字，有四个
• 所以两，三，四对应shape(2,3,4)