Question

我有numpy数组和两个python索引列表，它们的位置使数组元素增加一。 numpy是否有一些方法可以向量化此操作而不使用for循环？

我目前的执行缓慢：

a = np.zeros([4,5])
xs = [1,1,1,3]
ys = [2,2,3,0]

for x,y in zip(xs,ys): # how to do it in numpy way (efficiently)?
    a[x,y] += 1

print(a)

输出：

[[0. 0. 0. 0. 0.]
 [0. 0. 2. 1. 0.]
 [0. 0. 0. 0. 0.]
 [1. 0. 0. 0. 0.]]

Answer 1

pc.onnegotiationneeded = e => { if (pc.signalingState != "stable") return; .. }可以做到这一点，只需将两个索引作为单个2D数组/列表传递即可：

np.add.at

Answer 2

>>> a = np.zeros([4,5])
>>> xs = [1, 1, 1, 3]
>>> ys = [2, 2, 3, 0]
>>> a[[xs,ys]] += 1
>>> a
array([[ 0.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  1.,  1.,  0.],
       [ 0.,  0.,  0.,  0.,  0.],
       [ 1.,  0.,  0.,  0.,  0.]])

使用索引列表对numpy数组的元素执行操作

2 个答案: