Python的异常处理

常见错误

1.ZeroDivisionError(零除错误)

In [1]: 10/0---------------------------------------------------------------------------ZeroDivisionError                         Traceback (most recent call last)<ipython-input-1-242277fd9e32> in <module>()----> 1 10/0ZeroDivisionError: integer division or modulo by zero

2.IOError（文件错误）

In [2]: open('hello')---------------------------------------------------------------------------IOError                                   Traceback (most recent call last)<ipython-input-2-9834044a9e14> in <module>()----> 1 open('hello')IOError: [Errno 2] No such file or directory: 'hello'

3.SyntaxError（语法错误）

In [3]: for i in [1,2,3]  File "<ipython-input-3-ae71676907af>", line 1    for i in [1,2,3]                    ^SyntaxError: invalid syntax

4.IndexError（索引错误）

In [4]: t = (1,2,3)In [5]: t[3]---------------------------------------------------------------------------IndexError                                Traceback (most recent call last)<ipython-input-5-7d5cf04057c5> in <module>()----> 1 t[3]IndexError: tuple index out of range

5.KeyError（字典的key错误）

In [7]: d = {'name':'cici'}In [8]: d['name']Out[8]: 'cici'In [9]: d['age']---------------------------------------------------------------------------KeyError                                  Traceback (most recent call last)<ipython-input-9-39d5d5cefe61> in <module>()----> 1 d['age']KeyError: 'age'

6.NameError（名称错误）

In [10]: print a---------------------------------------------------------------------------NameError                                 Traceback (most recent call last)<ipython-input-10-9d7b17ad5387> in <module>()----> 1 print aNameError: name 'a' is not defined

错误排查

捕获异常

python异常处理机制: try….except…finally…
如果异常则执行except后的语句，没有异常则直接跳过，也可以用else来提示没有异常
finally后的语句是必然执行的，不管try后的语句是否正确

1.已知错误（跳过错误继续执行）

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"try:                        #用 try 来运行可能会出错的代码    print 'starting...'    li = [1, 2, 3]    print a    print li[3]except IndexError:              #用except来捕获错误,如果异常则执行    print 'index out of list'except NameError,e:             #捕获第二个错误，同时给错误赋值    print efinally:                        #必执行的    print 'ending..'

2.未知错误

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"try:    print 'starting...'    li = [1, 2, 3]    print a    print li[3]except BaseException,e:        #BaseException是一个错误的父类，表示所有错误    print 'Error:%s'%eelse:    print 'no error'finally:    print 'ending..'

3.没有异常的情况

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"try:    print 'starting...'    li = [1, 2, 3]    li.append(4)    a = 1    print a    print li[3]except IndexError:              #检测第一个错误    print 'index out of list'except NameError:               #检测第二个错误    print 'name is not defined'else:                           #没有异常时    print 'no error'finally:    print 'ending..'

4.函数排错
直接在主函数里排错，可以增强排错效率

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"def func1(s):    return func2(s) * 2def func2(s):    return 20 / sdef main():    try:        print func1('10')    except BaseException, e:        print emain()

5.错误日志

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"import logging                          ##导入日志模块logging.basicConfig(filename='err.log') #日志文件的定义def func1(s):    return func2(s) * 2def func2(s):    return 20 / sdef main():    try:        print func1('10')    except Exception as e:        logging.exception(e)    #所有错误导入到日志文件main()

抛出异常

except: 捕获异常
raise: 抛出异常

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"class MyError(BaseException):    passa = 1if a == 1:    raise  MyError

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"class MyError(BaseException):    passdef func1(s):    return func2(s) * 2def func2(s):    return 20 / int(s)def main():    try:        print func1('0')    except ZeroDivisionError as e:        raise MyError        #ZeroDivisionError 转换为 MyErrormain()

调试

print

用 print 把可能有问题的变量打印出来
用 print 最大的坏处是将来还得删掉它,运行结果也会包含很多垃圾信息。

断言

凡是用 print 来辅助查看的地方,都可以用断言(assert)来替代:;
如果断言失败, assert 语句本身就会抛出 AssertionError

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"def foo(s):    n = int(s)    return 10 / ndef main():    foo('0')assert foo(5) == 2print 'right'

Python 解释器执行时可以用 -O 参数来关闭 assert,把所有的 assert 语句当成 pass。
[kiosk@foundation32 date7.15]$ python -O error_4.py

logging

logging 不会抛出错误,而且可以输出到文件
logging.info() 就可以输出一段文本到日志文件中

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"import logginglogging.basicConfig(filename='logging.log',level=logging.INFO)def foo(s):    # 日志级别:debug,info,waring,error    n = int(s)    logging.info('n=%d'%n)    logging.warning('n=%d....warn'%n)    return 10 / ndef main():    foo('10')main()

pdb

[kiosk@foundation32 date7.15]$ python -m pdb error_3.py
n – next
p 变量 – 取出变量
q – 退出

#!/usr/bin/env python# coding:utf-8_author_ = "lvah"import pdbn1 =1n = int(n1)print npdb.set_trace()     #在可能出错的地方放一个 pdb.set_trace() ,就可以设置一个断点print 'hello'pdb.set_trace()s = 2print s

python …
程序会自动在 pdb.set_trace() 暂停并进入 pdb 调试环境, p 查看变量, c 继续运行