[转]Python文件操作

前言

这里的“文件”不单单指磁盘上的普通文件，也指代任何抽象层面上的文件。例如：通过URL打开一个Web页面“文件”，Unix系统下进程间通讯也是通过抽象的进程“文件”进行的。由于使用了统一的接口，从而统一了各种抽象类型及非抽象类型文件的操作方式。

文件操作的重要性无需多言，要想将计算机运算的结果以一定形式保存下来，文件是必须的。

一、内建函数open()和file()

open用法：    file_object = open(file_name, access_mode = 'r', buffering = -1)

• 返回的文件对象file_object是可迭代的，可以用for循环遍历

• file_name是要打开文件的文件名，非当前目录下的文件要指明路径

• access_mode是文件读取模式，默认为只读（r），这里的用法和C中的fopen是一样的

• buffering用来设定缓冲模式，默认值为-1，即使用系统默认缓冲模式

• 如果文件打开失败，则会返回IOError异常

1 2 3 4 5 6 7 8 9 10 11 12

#路径也可以写为 r'D:\Test.txt'                                                                                                                                                                                                                                                          >>> file_obj = open('D:\\Test.txt') #使用默认access_mode                                                                                                                                                                                                                                                                                              #和默认buffering值                                                                                                                                                                                                                                                          >>> for line in file_obj:  #file_obj可迭代                                                                                                                                                                                                                                                              print line                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            This is line1                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 I'm line2                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 Hello, World!

参数详细说明：

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

#文件对象访问模式                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              r  只读方式打开           rU  读方式打开，提供通用换行符支持                                                                                                                                                                                                                                                         w  以写方式打开           a   以追加模式打开                                                                                                                                                                                                                                                         r+ 以读写模式打开         w+  以读写模式打开                                                                                                                                                                                                                                                         rb 以二进制读模式打开     wb  以二进制写模式打开                                                                                                                                                                                                                                                         ab 以二进制追加模式打开   rb+ 以二进制读写模式打开                                                                                                                                                                                                                                                         wb+以二进制读写模式打开   ab+ 以二进制读写模式打开                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              #以r模式打开时，文件必须存在，否则引发错误                                                                                                                                                                                                                                                         #以w模式打开文件时，如果文件存在则清空，不存在则创建                                                                                                                                                                                                                                                         #以a模式打开时，如果文件存在，则从EOF位置追加，否则创建新文件                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              #buffering模式                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              0        不缓冲                                                                                                                                                                                                                                                         1        只缓冲一行                                                                                                                                                                                                                                                         value>1  缓冲区大小为value                                                                                                                                                                                                                                                         value<0  使用系统默认缓冲机制

内建工厂函数file()

file( )和open( )具有相同的功能和用法，可以任意替换。一般建议使用open()。

通用换行符支持（UNS）

在不同的系统平台上，换行符是不同的，例如unix下是\n，而windows下是\r\n，这就是为什么有时候你从网络上下载的txt小说不能自动换行的原因，这时候不要用记事本打开，使用windows自带的写字板或者浏览器打开再处理一下即可。

Python中为了解决这个问题，使用了通用换行符支持（Universal NEWLINE Support），使用带U模式的open打开文件时，Python会使用“\n”作为通用换行符，从而屏蔽了不同平台下的换行符差异。

这里不得不说的是，跨平台开发会遇到一些不可避免的问题，不同平台下的换行符差异及路径分隔符差异等等就是一个特例。幸运的是，Python的os模块提供了一些属性值以便于跨平台应用的开发：

os.linesep      （当前系统下，下同）用于在文件中分隔行的字符串

os.sep             用于分隔文件路径名的字符串

os.pathsep      用于分隔文件路径的字符串

os.curdir         当前工作目录的字符串名称，windows下是“.”

os.pardir        当前工作目录父目录的字符串名称，windows下是“..”

二、文件的内建方法、内建属性

输入

file_obj.read(size)  读取文件中的字节到字符串中，最多读取size个字节，如果没有给定size或者size为负数，文件将被读取直至末尾。这个函数不推荐使用。

file_obj.readline(size)  读取文件中size个字节，返回一个字符串。如果没有给定size或者size为负数则返回一行（包括行结束符）

1 2 3 4 5 6 7 8 9

>>> FILE2 = open('README.txt', 'rU')                                                                                                                                                                                                          >>> FILE2.readline()                                                                                                                                                                                                          'This is Python version 2.7.1\n' >>> FILE2.readline()                                                                                                                                                                                                          '============================\n' >>> FILE2.readline()                                                                                                                                                                                                          '\n' >>> FILE2.readline()                                                                                                                                                                                                          'Copyright (c) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010\n'

file_obj.readlines()  读取剩余的所有的行（文件指针不一定在开始位置！）并将其以一个字符串列表形式返回。此方法一次性读取文件所有的内容至内存中，适用于小型文件。对于大型文件，一般用迭代的方式读入内存，一会会讲到。

输出

file_obj.write()  write()方法和read及readline相反，它把含有文本数据或者二进制数据块的字符串写入到文件中。

file_obj.writelines()  和readlines()相反，它接收一个字符串列表作为参数并将其写入到文件中，每个字符串的行结束符不会被自动写入。需要的话，需要你自行添加。

移动文件指针

file_obj.seek( offset[,whence] )  seek()方法类似于C中的fseek，offset为偏移量，whence代表相对位置，是一个可选参数，默认为0（文件开头）。1代表当前位置，2代表文件末尾。

file_obj.tell( )  类似于C中的ftell，返回当前文件指针相对于文件头的位置。

1 2 3 4 5 6 7 8

>>> FILE = open('README.txt')                                                                                                                                                                    >>> FILE.readline()                                                                                                                                                                    'This is Python version 2.7.1\n' >>> FILE.tell()                                                                                                                                                                    30L >>> FILE.seek(-5, 1)  #从当前位置后退5个字节                                                                                                                                                                  >>> FILE.read(5)                                                                                                                                                                    '7.1\n='

文件迭代

1 2 3 4 5 6 7

#方法1 使用文件对象的readline()方法                                                                                                                                                            for lineItem in file_obj.readline():                                                                                                                                                                process lineItem                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       #方法2 直接迭代文件对象                                                                                                                                                            for lineItem in file_obj:                                                                                                                                                                process lineItem

注意，因文件对象本身就是可迭代的，所以使用方法2明显要优于方法1

采用迭代方式读取超大型文件或者网络流文件的好处是显而易见的，避免了一次性将大型文件读入内存所带来的负担（有时候甚至是不可能的，例如网络流文件）。对于小型文件还是一次性读入好些，可以尽快释放文件资源。

完美的终结

一切结束，千万不要忘记调用file_obj.close()以终结对文件的访问。如果你不这样显式的关闭文件，有可能丢失输出缓冲区内的数据。

一些其它方法

file_obj.fileno()  返回打开文件的文件描述符，这是一个整形，可以用于os模块的一些底层操作

file_obj.flush()   把输出缓冲区内的数据立即写入文件，调用close时会自动调用这个方法。

file_obj.isatty()  当文件是一个tty设备时，返回True。tty是字符型终端设备，例如老式的打印机以及操作系统中的终端（Terminal）程序。

file_obj.next()   返回文件的下一行，类似于readline方法，没有其它行时引发StopIteration异常

文件的内建属性

file.closed            如果文件已关闭，返回True

file.encodeing     文件所使用的编码

file.mode             文件打开时的模式

file.name             文件名

file.newlines        文件中用到的行分隔符，只有一种时为字符串，多于一种时是一个列表

file.softspace       这个属性程序员一般用不着，感兴趣的话请help(file.softspace)

三、标准文件、命令行参数

标准文件

一般来说，程序执行后，你可以访问3个标准文件。它们是标准输入（一般是键盘）、标准输出（到显示器的缓冲输出）、标准错误（到屏幕的非缓冲输出）。这些文件沿用了C中的命名：stdin，stdout，stderr。3个标准文件会被操作系统预先打开，因此，你只需要知道它们的文件句柄就可以随时访问它们。

1 2 3 4 5

>>> import sys                                                                  >>> sys.stdout.write('hello, This works like print')                                                                  hello, This works like print >>> sys.stderr.write('wahoo, This is a err I created!')                                                                  wahoo, This is a err I created!    #这行在IDLE中会显示为错误的红色

命令行参数

sys模块的argv属性是命令行参数组成的列表（Python中没有C中的argc），使用参照下例：

1 2 3 4 5 6 7

from sys import argv                                                                                                                                                                           numOfArg = len(argv)                                                                                                                                                                           print 'There are total %d arguments.' % numOfArg                                                        for num in range(numOfArg):                                                            print 'Arg No.%d:\t%s' % (num + 1, argv[num])

保存为arg.py，在windows的命令终端下运行效果如下：

四、文件系统

Python的os模块实现了操作文件系统的接口。这些操作包括遍历目录树，删除/重命名文件等。此外os.path模块可以实现一些针对路径名的操作。

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

#---------os模块的函数---------                                                                                                                  #文件处理                                     mkfifo()/mknod()     创建命名管道/文件系统节点                                     remove()/unlink()    删除文件                                     rename()/renames()   重命名文件                                     stat()               返回文件信息                                     symlink()            创建符号链接                                     utime()              更新时间戳                                     tmpfile()            创建并打开('w+b')一个新的临时文件                                     walk()               生成一个目录树下的所有文件名                                                                                                                  #目录/文件夹                                     chdir()/fcdir()      改变当前工作目录/通过一个文件描述符改变当前目录                                     chroot()             改变当前进程的根目录                                     listdir()            列出指定目录的文件                                     getcwd()/getcwdu()   返回当前工作目录/功能相同，但返回Unicode对象                                     mkdir()/mkdirs()     创建目录/创建多层目录                                     rmdir()/removedirs() 删除目录/删除多层目录                                                                                                                  #访问/权限                                     access()             检验权限模式                                     chmod()              改变权限模式                                     chown()/lchown()     改变owner和group ID/功能相同，但不会跟踪链接                                     umask()              设置默认权限模式                                                                                                                  #文件描述符操作                                     open()               底层的操作系统open（对于文件，使用标准的内建open)                                     read()/write()       根据文件描述符进行读写操作                                     dup()/dup2()         复制文件描述符号/功能相同，但复制到另一个文件描述符                                                                                                                  #设备号                                     makedev()            从major和minor设备号创建一个原始设备号                                     major()/minor()      从原始设备号获得major/minor设备号                                                                                                                  #---------os.path模块中的路径名访问函数---------                                                                                                                  #分隔                                     basename()           去掉目录路径，返回文件名                                     dirname()            去掉文件名，返回目录路径                                     join()               将分隔的部分组合成路径                                     split()              返回(dirname(), basename())元组                                     splitdrive()         返回(drivename, pathname)元组                                     splitext()           返回(filename, extension)元组                                                                                                                  #信息                                     getatime()           返回最近访问时间                                     getctime()           返回文件创建时间                                     getmtime()           返回最近文件修改时间                                     getsize()            返回文件大小                                                                                                                  #查询                                     exists()             指定路径（文件或者目录）是否存在                                     isabs()              指定路径是否为绝对路径                                     isdir()              指定路径是否存在且是一个目录                                     isfile()             指定路径是否存在且是一个文件                                     islink()             指定路径是否存在且是一个符号链接                                     ismount()            指定路径是否存在且是一个挂载点                                     samefile()           两个路径名是否指向同一个文件

除了以上列出的功能外，还可以利用os模块提供的功能进行系统进程管理以及进程间通讯，在后面的学习中会逐渐讲到。

五、一些相关模块

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

#文件相关模块                                                                                          base64           提供二进制字符串和文本字符串之间的编解码操作                             binascii         提供二进制和ASCII编码的二进制字符串间编解码操作                             bz2              访问BZ2格式的压缩文件                             csv              访问逗号分隔文件（csv）                             filecmp          用于比较目录和文件                             fileinput        提供多个文本文件的行迭代器                             getopt/optparse  提供了命令行参数的解析处理                             glob/fnmatch     提供了Unix样式的通配符匹配功能                             gzip/zlib        读写GNU zip文件（压缩需要zlib模块）                             shutil           提供高级文件访问功能                             c/StringIO       对字符串对象提供类文件接口                             tarfile          读写TAR归档文件，支持压缩文件                             tempfile         创建一个临时文件                             uu               uu格式的编解码                             zipfile          用于读取zip归档文件的工具

如果需要的话，可以浏览Python的参考文档查询这些模块的用法。或者import后再help。

自己可以试着写一些文本处理，特别是模版分析处理程序，试着把字符串处理和文件操作结合起来，可别小看这些功能，编码起来并不容易，需要十分的细致和耐心。当然，这是很好的编码能力锻炼方法。

要想对文本处理和文件操作有深入了解，如打包、压缩等，请参考《Python Cookbook》。

原文地址：http://greenlcat.diandian.com/post/2012-10-19/40039196726