对象拷贝
- 浅拷贝
- 深拷贝
浅拷贝与深拷贝
定义:进行浅拷贝时,先创建一个与拷贝对象1一样的类型对象2,然后再对对象2中的各个元素的引用拷贝并添加到对象2中。
Warning:不管是浅拷贝和深拷贝,只要拷贝的对象是非容器类型(比如字符串,数字和其他原子类型)就不能进行拷贝而是添加引用
拷贝方式:完全切片操作[:]、内建函数(list()、tuple()等)、copy模块的copy函数copy.copy()
>>> person = ['name', ['saving', 100]]
>>> hubby = person
>>> wife = person[:]
>>> [id(x) for x in person, hubby, wife]
[42440264L, 42440264L, 42452360L] #hubby与wife的身份不一样
>>> [id(x) for x in hubby ]
[36787064L, 37300232L]
>>> [id(x) for x in wife]
[36787064L, 37300232L]
>>> hubby, wife
(['name', ['saving', 50]], ['name', ['saving', 50]]) #1
>>> [id(x) for x in hubby ]
[36787064L, 37300232L]
>>> [id(x) for x in wife]
[36787064L, 37300232L]
为何注释1两个对象的100会变成50,这是因为无论改变hubby还是、wife的object[1][1]都是对person[1][1]的改变,没错就是引用,类似C语言中的指针。
深拷贝
定义:进行深拷贝时,先创建一个与拷贝对象1一样的类型对象2,然后再对对象2中的各个元素一模一样拷贝并添加到对象2中。
>>> import copy
>>> hubby = copy.deepcopy(person)
>>> wife, hubby
(['name', ['saving', 50]], ['name', ['saving', 50]])
>>> [id(x) for x in wife] #1
[36787064L, 423234242L]
>>> [id(x) for x in hubby]
[36787064L, 42504968L] #2
>>> wife[1][1] = 25
>>> wife, hubby
(['name', ['saving', 25]], ['name', ['saving', 50]]) #3
>>> [id(x) for x in hubby]
[36787064L, 42504968L] #5
>>> [id(x) for x in wife]
[36787064L, 37300232L] #6
>>> hubby[0] = 'Michael'
>>> wife[1] = 'ave'
>>> [id(x) for x in wife]
[36787064L, 42475728L] #7
>>> [id(x) for x in hubby]
[42476608L, 42504968L] #8
>>> a = 0
>>> import copy
>>> b = copy.deepcopy(a)
>>> id(a); id(b)
36987344L
36987344L
>>> a= [1,2,3,4,5]
>>> b = [1,2,3,4,5]
>>> id(a) ; id(b)
42507080L
42440072L
>>> b= (1,2,3,4,'hello')
>>> a = copy.deepcopy(b)
>>> id(a);id(b)
41897656L
41897656L
>>> a = [1,2,3,54,5,'hacker']
>>> b = copy.deepcopy(a)
>>> a
[1, 2, 3, 54, 5, 'hacker']
>>> id(a);id(b)
42440072L
42507464L
solve:
1和2和3注释:因为第二个元素是完全复制所以id不一样,不是同一个对象。至于第一个元素。以为其实字符串,所以请看上面的Warning。
7和8注释:因为第一个元素不是同一个对象。
最后两行:深拷贝对于内部元素只包含原子类型对象的元组只进行浅拷贝,对于列表则不会。
Summary:
两种拷贝都会创建一个新对象,通过id函数识别身份。
深拷贝对对象(除了原子类型)完全地拷贝,拷贝之后虽然;而浅拷贝只是对对象进行引用。