I don't know why the syntax is different when I take out a part of the list.

I wrote the following code:

dy[0][0:1] failed to retrieve the first number in the list, so dy_sum can be shaped once and dy_sum[0:1] can retrieve the first number in the list, but for some reason dx[0][0:1] must be .

I'm in trouble because I don't know the reason.Could someone explain it to me?

import numpy as np

D,N = 8,7
x = np.random.randn(1,D)
y=np.repeat(x,N,axis=0)

dy=np.random.randn(N,D)
dx=np.sum (dy,axis=0, keepdims=True)

dy_sum=np.sum(dy,axis=0)

print(dy_sum[0:1])
print(dx[0][0:1])