Python内置模块：copy

内置模块copy的存在主要为了深度拷贝，那么什么是深度拷贝呢？

深度拷贝会递归拷贝所有深度成员。

>>> import copy
>>> type(copy)
<type 'module'>
>>> dir(copy)
['Error', 'PyStringMap', '_EmptyClass', '__all__', '__builtins__', '__doc__', '__file__', '__name__', '__package__', '_copy_dispatch', '_copy_immutable', '_copy_inst', '_copy_with_constructor', '_copy_with_copy_method', '_deepcopy_atomic', '_deepcopy_dict', '_deepcopy_dispatch', '_deepcopy_inst', '_deepcopy_list', '_deepcopy_method', '_deepcopy_tuple', '_keep_alive', '_reconstruct', '_test', 'copy', 'deepcopy', 'dispatch_table', 'error', 'name', 't', 'weakref']
>>> help(copy.deepcopy)
Help on function deepcopy in module copy:

deepcopy(x, memo=None, _nil=[])
    Deep copy operation on arbitrary Python objects.

    See the module's __doc__ string for more info.

>>>

存在深度拷贝主要原因是可变数据类型可能会触碰到这个问题。

>>> l1 = [1, [2, 3], 4]
>>> l2 = l1[:]
>>> l1
[1, [2, 3], 4]
>>> l2
[1, [2, 3], 4]
>>> id(l1)
52764264
>>> id(l2)
52762824
>>> id(l1[1])
52786560
>>> id(l2[1])
52786560
>>> l1.append(5)
>>> l1
[1, [2, 3], 4, 5]
>>> l2
[1, [2, 3], 4]
>>> l1[1].append(100)
>>> l1
[1, [2, 3, 100], 4, 5]
>>> l2
[1, [2, 3, 100], 4]

上面代码是一个浅拷贝的例子，不经意间你的代码就会出现问题，如果用copy模块呢？

>>> import copy
>>> l1
[1, [2, 3, 100], 4, 5]
>>> l3 = copy.deepcopy(l1)
>>> id(l1)
52764264
>>> id(l3)
52786520
>>> id(l1[1])
52786560
>>> id(l3[1])
52788600
>>> l1[1].pop()
100
>>> l1
[1, [2, 3], 4, 5]
>>> l3
[1, [2, 3, 100], 4, 5]
>>>