例:
import numpy as np
a = np.random.randint(1,3,[4,4])
print(a)
print(a>1)
print(np.sum(a>1, axis=1))
print(np.where(np.sum(a>1, axis=1)))
print(np.where(np.sum(a>1, axis=1)==a.shape[1]))
print(np.where(np.sum(a>1, axis=1)==a.shape[1])[0])
输出结果:
# a:
[[2 1 2 2]
[2 2 2 2]
[2 1 2 1]
[2 2 1 2]]
# print(a>1)
[[ True False True True]
[ True True True True]
[ True False True False]
[ True True False True]]
# print(np.sum(a>1, axis=1))
[3 4 2 3]
# print(np.where(np.sum(a>1, axis=1))
(array([0, 1, 2, 3]),)
# print(np.where(np.sum(a>1, axis=1)==a.shape[1]))
(array([1]),)
# print(np.where(np.sum(a>1, axis=1)==a.shape[1])[0])
[1]
分析:
- a > 1 返回条件成立的a的bool值ndarray数组
- np.sum(a>1, axis=1)返回每一行的True个数数组,axis=1的意思是将求和结果映射到y轴
- np.where(np.sum(a>1, axis=1)返回的是传入的这个一维数组[3,4,2,3]中值不为0的索引数组
- np.where(np.sum(a>1, axis=1)==a.shape[1])表示的是返回[3,4,2,3]这个数组中值等于4的索引即1,这里a.shape[1]==4,那么该用法的功能就是返回上面的布尔数组中某行全为True即某行满足全部大于1的行索引。