python中的深拷贝和浅拷贝

copy.copy()中如果拷贝的是列表等可修改对象，则会拷贝最上一层，如果拷贝的最上层是元组等不可修改对象，则只是拷贝了对象的引用：

In [2]: a = [1, 2, 3]

In [3]: b = copy.copy(a)

In [4]: id(a)
Out[4]: 140568059574600

In [5]: id(b)
Out[5]: 140568059649992

In [6]: a = (1, 2, 3)

In [7]: b = copy.copy(a)

In [8]: id(a)
Out[8]: 140568101534096

In [9]: id(b)
Out[9]: 140568101534096

 

copy.deepcopy()中如果拷贝的是不可变对象，依然是拷贝对象的引用：

In [2]: a = (1, 2, 3)

In [3]: b = copy.deepcopy(a)

In [4]: id(a)
Out[4]: 140313481712336

In [5]: id(b)
Out[5]: 140313481712336

 

而如果嵌套内部包含可修改和不可修改对象，则会逐级拷贝所有的可修改对象并只拷贝所有的不可修改对象的引用：

In [6]: a = [1, 3]

In [7]: b = (2, 4)

In [8]: c = [a, b]

In [9]: d = copy.deepcopy(c)

In [10]: id(c)
Out[10]: 140313510332424

In [11]: id(d)
Out[11]: 140313468881672

In [12]: id(c[1])
Out[12]: 140313468353096

In [13]: id(d[1])
Out[13]: 140313468353096

In [14]: id(c[0])
Out[14]: 140313468365640

In [15]: id(d[0])
Out[15]: 140313468161608

如果内层包含有可修改对象，则无论外层为可修改对象还是不可修改对象，对外层对象都会进行拷贝，而不是仅对外层对象的引用进行拷贝：

In [16]: a = [1, 3]

In [17]: b = (2, 4)

In [18]: c = (a, b)

In [19]: d = copy.deepcopy(c)

In [20]: id(c)
Out[20]: 140313510401352

In [21]: id(d)
Out[21]: 140313482148872

In [22]: id(c[0])
Out[22]: 140313482149192

In [23]: id(d[0])
Out[23]: 140313468180744

In [24]: id(c[1])
Out[24]: 140313479821256

In [25]: id(d[1])
Out[25]: 140313479821256

In [26]: e = (c, (1, 3))

In [27]: f = copy.deepcopy(e)

In [28]: id(e)
Out[28]: 140313510249160

In [29]: id(f)
Out[29]: 140313468328072

In [30]: id(e[0])
Out[30]: 140313510401352

In [31]: id(f[0])
Out[31]: 140313479822280

In [32]: id(e[1])
Out[32]: 140313510370888

In [33]: id(f[1])
Out[33]: 140313510370888

 

切片也是浅拷贝的一种：

In [42]: a = (1, 3, 5)

In [43]: b = a[:]

In [44]: id(a)
Out[44]: 140313506521160

In [45]: id(b)
Out[45]: 140313506521160

In [46]: a = [1, 3, 5]

In [47]: b = a[:]

In [48]: id(a)
Out[48]: 140313468221256

In [49]: id(b)
Out[49]: 140313479813576

In [50]: id(a[0])
Out[50]: 10914496

In [51]: id(b[0])
Out[51]: 10914496

In [52]: a = [1, 3, [2, 4]]

In [53]: b = a[:]

In [54]: id(a)
Out[54]: 140313510367816

In [55]: id(b)
Out[55]: 140313468221256

In [56]: id(a[2])
Out[56]: 140313479811016

In [57]: id(b[2])
Out[57]: 140313479811016