使用索引列表對numpy數組的元素執行操作

Question

我有numpy數組和兩個python索引列表，它們的位置將數組元素增加一。 numpy是否有一些方法可以向量化此操作而不使用for循環？

我目前執行緩慢：

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)

輸出：

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

Answer 1

np.add.at可以做到這一點，只需將兩個索引作為單個2D數組/列表傳遞：

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

np.add.at(a, [xs, ys], 1) # in-place
print(a)

array([[0., 0., 0., 0., 0.],
       [0., 0., 2., 1., 0.],
       [0., 0., 0., 0., 0.],
       [1., 0., 0., 0., 0.]])

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.]])