Posted in Python onJune 04, 2014
一、不要使用可变对象作为函数默认值
In [1]: def append_to_list(value, def_list=[]): ...: def_list.append(value) ...: return def_list ...: In [2]: my_list = append_to_list(1) In [3]: my_list Out[3]: [1] In [4]: my_other_list = append_to_list(2) In [5]: my_other_list Out[5]: [1, 2] # 看到了吧,其实我们本来只想生成[2] 但是却把第一次运行的效果页带了进来 In [6]: import time In [7]: def report_arg(my_default=time.time()): ...: print(my_default) ...: In [8]: report_arg() # 第一次执行 1399562371.32 In [9]: time.sleep(2) # 隔了2秒 In [10]: report_arg() 1399562371.32 # 时间竟然没有变
这2个例子说明了什么? 字典,集合,列表等等对象是不适合作为函数默认值的. 因为这个默认值实在函数建立的时候就生成了, 每次调用都是用了这个对象的”缓存”. 我在上段时间的分享python高级编程也说到了这个问题,这个是实际开发遇到的问题,好好检查你学过的代码, 也许只是问题没有暴露
可以这样改:
def append_to_list(element, to=None): if to is None: to = [] to.append(element) return to
二、生成器不保留迭代过后的结果
In [12]: gen = (i for i in range(5)) In [13]: 2 in gen Out[13]: True In [14]: 3 in gen Out[14]: True In [15]: 1 in gen Out[15]: False # 1为什么不在gen里面了? 因为调用1->2,这个时候1已经不在迭代器里面了,被按需生成过了 In [20]: gen = (i for i in range(5)) In [21]: a_list = list(gen) # 可以转化成列表,当然a_tuple = tuple(gen) 也可以 In [22]: 2 in a_list Out[22]: True In [23]: 3 in a_list Out[23]: True In [24]: 1 in a_list # 就算循环过,值还在 Out[24]: True
三、lambda在闭包中会保存局部变量
In [29]: my_list = [lambda: i for i in range(5)] In [30]: for l in my_list: ....: print(l()) ....: 4 4 4 4 4
这个问题还是上面说的python高级编程中说过具体原因. 其实就是当我赋值给my_list的时候,lambda表达式就执行了i会循环,直到 i =4,i会保留
但是可以用生成器
In [31]: my_gen = (lambda: n for n in range(5)) In [32]: for l in my_gen: ....: print(l()) ....: 0 1 2 3 4
也可以坚持用list:
In [33]: my_list = [lambda x=i: x for i in range(5)] # 看我给每个lambda表达式赋了默认值 In [34]: for l in my_list: ....: print(l()) ....: 0 1 2 3 4
有点不好懂是吧,在看看python的另外一个魔法:
In [35]: def groupby(items, size): ....: return zip(*[iter(items)]*size) ....: In [36]: groupby(range(9), 3) Out[36]: [(0, 1, 2), (3, 4, 5), (6, 7, 8)]
一个分组的函数,看起来很不好懂,对吧? 我们来解析下这里
In [39]: [iter(items)]*3 Out[39]: [<listiterator at 0x10e155fd0>, <listiterator at 0x10e155fd0>, <listiterator at 0x10e155fd0>] # 看到了吧, 其实就是把items变成可迭代的, 重复三回(同一个对象哦), 但是别忘了,每次都.next(), 所以起到了分组的作用 In [40]: [lambda x=i: x for i in range(5)] Out[40]: [<function __main__.<lambda>>, <function __main__.<lambda>>, <function __main__.<lambda>>, <function __main__.<lambda>>, <function __main__.<lambda>>] # 看懂了吗?
四、在循环中修改列表项
In [44]: a = [1, 2, 3, 4, 5] In [45]: for i in a: ....: if not i % 2: ....: a.remove(i) ....: In [46]: a Out[46]: [1, 3, 5] # 没有问题 In [50]: b = [2, 4, 5, 6] In [51]: for i in b: ....: if not i % 2: ....: b.remove(i) ....: In [52]: b Out[52]: [4, 5] # 本来我想要的结果应该是去除偶数的列表
思考一下,为什么 ? 是因为你对列表的remove,影响了它的index
In [53]: b = [2, 4, 5, 6] In [54]: for index, item in enumerate(b): ....: print(index, item) ....: if not item % 2: ....: b.remove(item) ....: (0, 2) # 这里没有问题 2被删除了 (1, 5) # 因为2被删除目前的列表是[4, 5, 6], 所以索引list[1]直接去找5, 忽略了4 (2, 6)
五、IndexError - 列表取值超出了他的索引数
In [55]: my_list = [1, 2, 3, 4, 5] In [56]: my_list[5] # 根本没有这个元素 --------------------------------------------------------------------------- IndexError Traceback (most recent call last) <ipython-input-56-037d00de8360> in <module>() ----> 1 my_list[5] IndexError: list index out of range # 抛异常了 In [57]: my_list[5:] # 但是可以这样, 一定要注意, 用好了是trick,用错了就是坑啊 Out[57]: []
六、重用全局变量
In [58]: def my_func(): ....: print(var) # 我可以先调用一个未定义的变量 ....: In [59]: var = 'global' # 后赋值 In [60]: my_func() # 反正只要调用函数时候变量被定义了就可以了 global In [61]: def my_func(): ....: var = 'locally changed' ....: In [62]: var = 'global' In [63]: my_func() In [64]: print(var) global # 局部变量没有影响到全局变量 In [65]: def my_func(): ....: print(var) # 虽然你全局设置这个变量, 但是局部变量有同名的, python以为你忘了定义本地变量了 ....: var = 'locally changed' ....: In [66]: var = 'global' In [67]: my_func() --------------------------------------------------------------------------- UnboundLocalError Traceback (most recent call last) <ipython-input-67-d82eda95de40> in <module>() ----> 1 my_func() <ipython-input-65-0ad11d690936> in my_func() 1 def my_func(): ----> 2 print(var) 3 var = 'locally changed' 4 UnboundLocalError: local variable 'var' referenced before assignment In [68]: def my_func(): ....: global var # 这个时候得加全局了 ....: print(var) # 这样就能正常使用 ....: var = 'locally changed' ....: In [69]: var = 'global' In [70]: In [70]: my_func() global In [71]: print(var) locally changed # 但是使用了global就改变了全局变量
七、拷贝可变对象
In [72]: my_list1 = [[1, 2, 3]] * 2 In [73]: my_list1 Out[73]: [[1, 2, 3], [1, 2, 3]] In [74]: my_list1[1][0] = 'a' # 我只修改子列表中的一项 In [75]: my_list1 Out[75]: [['a', 2, 3], ['a', 2, 3]] # 但是都影响到了 In [76]: my_list2 = [[1, 2, 3] for i in range(2)] # 用这种循环生成不同对象的方法就不影响了 In [77]: my_list2[1][0] = 'a' In [78]: my_list2 Out[78]: [[1, 2, 3], ['a', 2, 3]]
八、python多继承(C3)
In [1]: class A(object): ...: def foo(self): ...: print("class A") ...: In [2]: class B(object): ...: def foo(self): ...: print("class B") ...: In [3]: class C(A, B): ...: pass ...: In [4]: C().foo() class A # 例子很好懂, C继承了A和B,从左到右,发现A有foo方法,返回了
看起来都是很简单, 有次序的从底向上,从前向后找,找到就返回. 再看例子:
In [5]: class A(object): ...: def foo(self): ...: print("class A") ...: In [6]: class B(A): ...: pass ...: In [7]: class C(A): ...: def foo(self): ...: print("class C") ...: In [8]: class D(B,C): ...: pass ...: In [9]: D().foo() class C # ? 按道理, 顺序是 D->B->A,为什么找到了C哪去了
这也就涉及了MRO(Method Resolution Order):
In [10]: D.__mro__ Out[10]: (__main__.D, __main__.B, __main__.C, __main__.A, object)
简单的理解其实就是新式类是广度优先了, D->B, 但是发现C也是继承A,就先找C,最后再去找A
九、列表的+和+=, append和extend
In [17]: print('ID:', id(a_list)) ('ID:', 4481323592) In [18]: a_list += [1] In [19]: print('ID (+=):', id(a_list)) ('ID (+=):', 4481323592) # 使用+= 还是在原来的列表上操作 In [20]: a_list = a_list + [2] In [21]: print('ID (list = list + ...):', id(a_list)) ('ID (list = list + ...):', 4481293056) # 简单的+其实已经改变了原有列表 In [28]: a_list = [] In [29]: id(a_list) Out[29]: 4481326976 In [30]: a_list.append(1) In [31]: id(a_list) Out[31]: 4481326976 # append 是在原有列表添加 In [32]: a_list.extend([2]) In [33]: id(a_list) Out[33]: 4481326976 # extend 也是在原有列表上添加
十、datetime也有布尔值
这是一个坑
In [34]: import datetime In [35]: print('"datetime.time(0,0,0)" (Midnight) ->', bool(datetime.time(0,0,0))) ('"datetime.time(0,0,0)" (Midnight) ->', False) In [36]: print('"datetime.time(1,0,0)" (1 am) ->', bool(datetime.time(1,0,0))) ('"datetime.time(1,0,0)" (1 am) ->', True)
十一、'==' 和 is 的区别
我的理解是”is”是判断2个对象的身份, ==是判断2个对象的值
In [37]: a = 1 In [38]: b = 1 In [39]: print('a is b', bool(a is b)) ('a is b', True) In [40]: c = 999 In [41]: d = 999 In [42]: print('c is d', bool(c is d)) ('c is d', False) # 原因是python的内存管理,缓存了-5 - 256的对象 In [43]: print('256 is 257-1', 256 is 257-1) ('256 is 257-1', True) In [44]: print('257 is 258-1', 257 is 258 - 1) ('257 is 258-1', False) In [45]: print('-5 is -6+1', -5 is -6+1) ('-5 is -6+1', True) In [46]: print('-7 is -6-1', -7 is -6-1) ('-7 is -6-1', False) In [47]: a = 'hello world!' In [48]: b = 'hello world!' In [49]: print('a is b,', a is b) ('a is b,', False) # 很明显 他们没有被缓存,这是2个字段串的对象 In [50]: print('a == b,', a == b) ('a == b,', True) # 但他们的值相同 # But, 有个特例 In [51]: a = float('nan') In [52]: print('a is a,', a is a) ('a is a,', True) In [53]: print('a == a,', a == a) ('a == a,', False) # 亮瞎我眼睛了~
十二、浅拷贝和深拷贝
我们在实际开发中都可以向对某列表的对象做修改,但是可能不希望改动原来的列表. 浅拷贝只拷贝父对象,深拷贝还会拷贝对象的内部的子对象
In [65]: list1 = [1, 2] In [66]: list2 = list1 # 就是个引用, 你操作list2,其实list1的结果也会变 In [67]: list3 = list1[:] In [69]: import copy In [70]: list4 = copy.copy(list1) # 他和list3一样 都是浅拷贝 In [71]: id(list1), id(list2), id(list3), id(list4) Out[71]: (4480620232, 4480620232, 4479667880, 4494894720) In [72]: list2[0] = 3 In [73]: print('list1:', list1) ('list1:', [3, 2]) In [74]: list3[0] = 4 In [75]: list4[1] = 4 In [76]: print('list1:', list1) ('list1:', [3, 2]) # 对list3和list4操作都没有对list1有影响 # 再看看深拷贝和浅拷贝的区别 In [88]: from copy import copy, deepcopy In [89]: list1 = [[1], [2]] In [90]: list2 = copy(list1) # 还是浅拷贝 In [91]: list3 = deepcopy(list1) # 深拷贝 In [92]: id(list1), id(list2), id(list3) Out[92]: (4494896592, 4495349160, 4494896088) In [93]: list2[0][0] = 3 In [94]: print('list1:', list1) ('list1:', [[3], [2]]) # 看到了吧 假如你操作其子对象 还是和引用一样 影响了源 In [95]: list3[0][0] = 5 In [96]: print('list1:', list1) ('list1:', [[3], [2]]) # 深拷贝就不会影响
十三、bool其实是int的子类
In [97]: isinstance(True, int) Out[97]: True In [98]: True + True Out[98]: 2 In [99]: 3 * True + True Out[99]: 4 In [100]: 3 * True - False Out[100]: 3 In [104]: True << 10 Out[104]: 1024
十五、元组是不是真的不可变?
In [111]: tup = ([],) In [112]: tup[0] += [1] --------------------------------------------------------------------------- TypeError Traceback (most recent call last) <ipython-input-112-d4f292cf35de> in <module>() ----> 1 tup[0] += [1] TypeError: 'tuple' object does not support item assignment In [113]: tup Out[113]: ([1],) # 我靠 又是亮瞎我眼睛,明明抛了异常 还能修改? In [114]: tup = ([],) In [115]: tup[0].extend([1]) In [116]: tup[0] Out[116]: [1] # 好吧,我有点看明白了, 虽然我不能直接操作元组,但是不能阻止我操作元组中可变的子对象(list)
这里有个不错的解释Python's += Is Weird, Part II :
In [117]: my_tup = (1,) In [118]: my_tup += (4,) In [119]: my_tup = my_tup + (5,) In [120]: my_tup Out[120]: (1, 4, 5) # ? 嗯 不是不能操作元组嘛? In [121]: my_tup = (1,) In [122]: print(id(my_tup)) 4481317904 In [123]: my_tup += (4,) In [124]: print(id(my_tup)) 4480606864 # 操作的不是原来的元组 所以可以 In [125]: my_tup = my_tup + (5,) In [126]: print(id(my_tup)) 4474234912
十六、python没有私有方法/变量? 但是可以有”伪”的
In [127]: class my_class(object^E): .....: def public_method(self): .....: print('Hello public world!') .....: def __private_method(self): # 私有以双下划线开头 .....: print('Hello private world!') .....: def call_private_method_in_class(self): .....: self.__private_method() In [132]: my_instance = my_class() In [133]: my_instance.public_method() Hello public world! # 普通方法 In [134]: my_instance._my_class__private_method() Hello private world! # 私有的可以加"_ + 类名字 + 私有方法名字” In [135]: my_instance.call_private_method_in_class() Hello private world! # 还可以通过类提供的公有接口内部访问 In [136]: my_instance._my_class__private_variable Out[136]: 1
十七、异常处理加else
In [150]: try: .....: print('third element:', a_list[2]) .....: except IndexError: .....: print('raised IndexError') .....: else: .....: print('no error in try-block') # 只有在try里面没有异常的时候才会执行else里面的表达式 .....: raised IndexError # 抛异常了 没完全完成 In [153]: i = 0 In [154]: while i < 2: .....: print(i) .....: i += 1 .....: else: .....: print('in else') .....: 0 1 in else # while也支持哦~ In [155]: i = 0 In [156]: while i < 2: .....: print(i) .....: i += 1 .....: break .....: else: .....: print('completed while-loop') .....: 0 # 被break了 没有完全执行完 就不执行else里面的了 In [158]: for i in range(2): .....: print(i) .....: else: .....: print('completed for-loop') .....: 0 1 completed for-loop In [159]: for i in range(2): .....: print(i) .....: break .....: else: .....: print('completed for-loop') .....: 0 # 也是因为break了
Python程序员鲜为人知但你应该知道的17个问题
声明:登载此文出于传递更多信息之目的,并不意味着赞同其观点或证实其描述。
Reply on: @reply_date@
@reply_contents@