numpy sum axis解释

 关于a.sum(axis=1)，网上有很多博客解释，但都说的不明白。我在stackoverflow上发现一个解释[1]，很容易记忆，录在这里。
 code.py

import  numpy as np
if __name__ == '__main__':
    a = np.arange(30).reshape(2, 3, 5)
    print(a)
    print(a[0,:,:]+a[1,:,:])
    print("\n")
    print(a.sum(axis=0))
    print("\n\n")
    print(a[:,0,:]+a[:,1,:]+a[:,2,:])
    print("\n")
    print(a.sum(axis=1))
    
    print("\n\n")
    print(a[:,:,0]+a[:,:,1]+a[:,:,2]+a[:,:,3]+a[:,:,4])
    print("\n")
    print(a.sum(axis=2))

 a.sum(axis=0)的结果同a[0,:,:]+a[1,:,:]相同。
 a.sum(axis=1)的结果同a[:,0,:]+a[:,1,:]+a[:,2,:]相同。
 a.sum(axis=2)的结果同a[:,:,0]+a[:,:,1]+a[:,:,2]+a[:,:,3]+a[:,:,4]相同。

Reference:
[1] How does NumPy Sum (with axis) work?