介绍引用传递

python只允许使用引用传递， 不存在其他语言中的值传递。引用传递即引用内存地址， 无论是赋值还是函数调用。

不可变对象

不可变对象指向内存地址的值不能够被改变， 这些对象包含int,string,float,tuple。

i = 1
j = i

print('id is %s for i' % id(i))
print('id is %s for j' % id(j))
print(i, j)

j = j + 1

print('id is %s for i' % id(i))
print('id is %s for j' % id(j))
print(i, j)
#result
id is 11061440 for i
id is 11061440 for j
1 1
id is 11061440 for i
id is 11061472 for j
1 2

• 最初 i 赋值给 j 时是引用传递， 它们的内存地址都是11061440， 即 i 和 j 指向同一个内存地址
• 后面当对象 j 被改变时， 由于所指向的内存中的值不能被改变， 所以会复制一份值到新的地址再处理计算(+1)， 然后对象 j 指向新的地址11061472
  不可变对象

可变对象

可变对象指向内存地址的值直接被改变， 这些对象包含list,dict,set。

a = [1]
b = a

print('id is %s for a' % id(a))
print('id is %s for b' % id(b))
print(a, b)

b[0] = 0

print('id is %s for a' % id(a))
print('id is %s for b' % id(b))
print(a, b)
#result
id is 140198854737480 for a
id is 140198854737480 for b
[1] [1]
id is 140198854737480 for a
id is 140198854737480 for b
[0] [0]

• 最初 a 赋值给 b 时是引用传递， 它们的内存地址都是140198854737480， 即 a 和 b 指向同一个内存地址
• 后面当对象 b 被改变时， 由于所指向的内存中的值直接被改变， 发现改变b的同时a也被改变了， 但是 a 和 b 还是指向同一个内存地址
  可变对象

函数的参数传递也分可变和不可变对象的

def test_para(a, b):
    a += 1
    b[0] = 0
    print('========1111111111==============')
    print('a is %s and the id is %s.' % (a, id(a)))
    print('b is %s and the id is %s.' % (b, id(b)))

a = 1
b = [1]
print('a is %s and the id is %s.' % (a, id(a)))
print('b is %s and the id is %s.' % (b, id(b)))
test_para(a, b)
print('========22222222222==============')
print('a is %s and the id is %s.' % (a, id(a)))
print('b is %s and the id is %s.' % (b, id(b)))

#result
a is 1 and the id is 11061440.
b is [1] and the id is 139909172055624.
========1111111111==============
a is 2 and the id is 11061472.
b is [0] and the id is 139909172055624.
========22222222222==============
a is 1 and the id is 11061440.
b is [0] and the id is 139909172055624.

• 当传过来的是可变对象，函数内部修改会影响函数外部的可变对象。
• 当传过来的是不可变对象，函数内部修改不会影响函数外部的不可变对象，因为修改的时候会先复制一份再修改

def test_para(b=[]):
    b += [0]
    print('b is %s and the id is %s.' % (b, id(b)))

test_para()
test_para()
test_para()
test_para([1, 2, 3])

#result
b is [0] and the id is 140394154311240.
b is [0, 0] and the id is 140394154311240.
b is [0, 0, 0] and the id is 140394154311240.
b is [1, 2, 3, 0] and the id is 140394177812936.

• Python 解释器执行 def 语句时， 会创建一个函数对象， 且只会有一个
• 默认参数 b 就已经计算出来， 指向了 [] 空列表
• 第一次调用test_para()时， 往b中添加一个值为1的元素
• 第二次，第三次调用test_para()时， 继续往b中添加一个值为1的元素， 用了同一个对象b， 指向同一个内存地址
• 当调用test_para([1, 2, 3])时， b 被重新赋值， b指向了[1, 2, 3]的新地址

list中的注意事项

• 对象被重新赋值
  a = a + [3] 或者 a = [3] 都是重新被赋值， 但是 a += [3]不是的，相当于a.extend([3])

a = b = [1, 2]
print(a, id(a), b, id(b))

a = a + [3]
print(a, id(a), b, id(b))
#result
[1, 2] 140607382289992 [1, 2] 140607382289992
[1, 2, 3] 140607405791752 [1, 2] 140607382289992

a = b = [1, 2]
print(a, id(a), b, id(b))

a += [3]
print(a, id(a), b, id(b))
#result
[1, 2] 140479908183624 [1, 2] 140479908183624
[1, 2, 3] 140479908183624 [1, 2, 3] 140479908183624

a = b = [1, 2]
print(a, id(a), b, id(b))

a = [3]
print(a, id(a), b, id(b))
#result
[1, 2] 139779074858568 [1, 2] 139779074858568
[3] 139779098360264 [1, 2] 139779074858568

a = [1, 2]
s = [a] * 2
print(s, id(s[0]), id(s[1]))
a.append(3)
print(s, id(s[0]), id(s[1]))
a[0] = 0
print(s, id(s[0]), id(s[1]))
s[0] = [0]
print(s, id(s[0]), id(s[1]))
#result
[[1, 2], [1, 2]] 139773557825096 139773557825096
[[1, 2, 3], [1, 2, 3]] 139773557825096 139773557825096
[[0, 2, 3], [0, 2, 3]] 139773557825096 139773557825096
[[0], [0, 2, 3]] 139773581326792 139773557825096

• 如何复制list的值，但不指向同一个地址
  一阶列表和高阶列表有很大区别， 一阶列表可以使用a[:], list(a), a.copy(), copy.copy(a), copy.deepcopy(a) 等方法拷贝一份， 而高阶列表只能使用copy.deepcopy(a)深度拷贝一份

import copy

a = [0, 1]
b = 3
num = [a, b]
print('num is %s and id is %s\n' % (num, id(num)))

c = num[:]
d = list(num)
e = num.copy()
f = copy.copy(num)
g = copy.deepcopy(num)

print('c is %s and id is %s\n' % (c, id(c)),
      'd is %s and id is %s\n' % (d, id(d)),
      'e is %s and id is %s\n' % (e, id(e)),
      'f is %s and id is %s\n' % (f, id(f)),
      'g is %s and id is %s\n' % (g, id(g))
      )
print('-------------------------------------------')
a.append(2)
print('num is %s and id is %s\n' % (num, id(num)))
print('c is %s and id is %s\n' % (c, id(c)),
      'd is %s and id is %s\n' % (d, id(d)),
      'e is %s and id is %s\n' % (e, id(e)),
      'f is %s and id is %s\n' % (f, id(f)),
      'g is %s and id is %s\n' % (g, id(g))
      )

#result
num is [[0, 1], 3] and id is 140319229340360

c is [[0, 1], 3] and id is 140319229131848
 d is [[0, 1], 3] and id is 140319229219464
 e is [[0, 1], 3] and id is 140319229219016
 f is [[0, 1], 3] and id is 140319229219784
 g is [[0, 1], 3] and id is 140319229220232

-------------------------------------------
num is [[0, 1, 2], 3] and id is 140319229340360

c is [[0, 1, 2], 3] and id is 140319229131848
 d is [[0, 1, 2], 3] and id is 140319229219464
 e is [[0, 1, 2], 3] and id is 140319229219016
 f is [[0, 1, 2], 3] and id is 140319229219784
 g is [[0, 1], 3] and id is 140319229220232