python现象记录

指针效应

即使经过序列化和反序列化，指针的效应仍然存在：

a = [[1,1], [1,1]]
a = a + a
a[0][0] = 999
print(a)
import _pickle as cpickle
with open("./test.pkl", 'wb') as f:
    cpickle.dump(a, f)
with open("./test.pkl", 'rb') as f:
    temp = cpickle.load(f)
temp[0][0] = 123
print(temp)
temp[0] = 123
print(temp)

结果如下：

[[999, 1], [1, 1], [999, 1], [1, 1]]
[[123, 1], [1, 1], [123, 1], [1, 1]]
[123, [1, 1], [123, 1], [1, 1]]

而构造成Tensor后，则不会出现此现象：

a = [[1,1], [1,1]]
a = torch.tensor(a+a)
print(a)
a[0][0] = 123
print(a)

结果如下：

tensor([[1, 1],
        [1, 1],
        [1, 1],
        [1, 1]])
tensor([[9, 1],
        [1, 1],
        [1, 1],
        [1, 1]])

string的map并不慢于int

和c++的map不同，python中string插入到集合中并不会显著慢于int的插入，如图：