File operations in Python

Applications that we write in Python or other languages need to be saved on the hard drive if we want to keep the data permanently, which involves the application operating hardware, and it is well known that the application is not able to manipulate the hardware directly. The operating system encapsulates complex hardware operations into a simple interface for use by the user/application, where the file is the operating system provided to the application to manipulate the virtual concept of the hard disk, and the user or application can save its own data permanently by manipulating the file.

I. Basic procedures for file operations

With the concept of a file, we no longer have to consider the details of the operation of the hard disk, just the process of manipulating the files:

#1. Open the file, get the file handle and assign a value to a variable f=open (' A.txt ', ' R ', encoding= ' Utf-8 ') #默认打开模式就为r #. Manipulate the file through a handle data=f.read () #3. Close File F.close ()

Considerations for closing files

Open a file that contains two parts of the resource: an operating system-level Open file + Application variable. When a file is completed, the two parts of the file must be recovered in a non-landed manner: 1, F.close () #回收操作系统级打开的文件2, Del f #回收应用程序级的变量其中del f must occur after f.close (), Otherwise it will cause the operating system to open the file is not closed, the use of resources, and Python automatic garbage collection mechanism determines that we do not have to consider Del F, which requires us, after the completion of the file, we must remember F.close () although I say so, But many students will not forget the face of F.close (), for those who do not long brain, we recommend a fool mode of operation: Use with the keyword to help us manage the context with open (' A.txt ', ' W ') as F:pass with open (' A.txt ', ' R ') as Read_f,open (' B.txt ', ' W ') as Write_f:data=read_f.read () write_f.write (data) note

Two. File encoding

F=open (...) is opened by the operating system file, then if we do not specify the encoding for open, then the default encoding of the opening file is obviously the operating system, the operating system will use its own default encoding to open the file, under Windows is GBK, under Linux is Utf-8

#这就用到了上节课讲的字符编码的知识: To ensure that it is not garbled, how the file stored in what way, it will open in what way. F=open (' A.txt ', ' R ', encoding= ' utf-8 ')

Three. File Open mode

#1. The mode of opening the file has (default is text mode): R, read-only mode "default mode, the file must exist, does not exist" throw exception "W, write-only mode" is not readable, does not exist, it is empty content "A, append write mode only" unreadable, does not exist, then only append the content "#2. For non-text files, we can only use B mode, "B" means to operate in bytes (and all files are stored in bytes, using this mode regardless of the text file character encoding, picture file JGP format, video file avi format) RB Wbab Note: When opened in B mode, Read to the content is a byte type, write also need to provide byte type, can not specify the encoding # #, ' + ' mode (is added a function) r+, read and write "readable, can write" w+, write read "writable, readable" A +, write read "writable, readable" #4, to bytes type of operation of Read and write, read, Write read mode r+b, read and write "readable, can write" w+b, write read "writable, readable" a+b, write read "writable, readable"

Four. File operation method Common file Operation method

Read (3
1. When the file is opened in text mode, the representation reads 3 characters
2. When the file is opened in B mode, the delegate reads 3 bytes
The rest of the files within the cursor movement are in bytes such as: seek,tell,truncate
Attention:
1. Seek has three modes of movement 0,1,2, of which 1 and 2 must be performed in B mode, but regardless of which mode is moved in bytes units
2. Truncate is a truncated file, so the file must be opened in a way that can be written, but not with W or w+, etc., because the file is directly emptied, so truncate to r+ or a or a + and other modes to test the effect.

All action methods.

Class textiowrapper (_textiobase):     "" "    character and  line based layer over a BufferedIOBase object, buffer.         encoding gives the name of the encoding  that the stream will be    decoded or encoded  With. it defaults to locale.getpreferredencoding (False) .         errors determines the strictness of encoding and decoding   (See    help (codecs. CODEC)  or the documentation for codecs.register)  and     defaults to  "Strict".         newline controls how  line endings are handled. It can be None,  ',     ' \ n ',  ' \ R ', and  ' \ r \ n ' .  it works as follows:         * On input, if newline is None,  universal newlines mode is      enabled. lines in  the input can end in  ' \ n ',  ' \ R ', or  ' \ r \ n ', and       these are translated into  ' \ n '  before being returned  to the      caller. If it is  ',  universal  newline mode is enabled, but line      endings  are returned to the caller untranslated. if it has any  of      the other legal values, input lines  Are only terminated by the given      string, and the line  ending is returned to the caller untranslated.         * On output, if newline is None, any  ' \ n '   Characters written are      translated to the system  default line separator, os.linesep. if      newline  is  '  or  ' \ n ', no translation takes place. if newline  is any      of the other legal values, any  ' \ N '  characters written are translated      to the  Given string.        if line_buffering is true,  a call to fLush is implied when a call to    write contains a  newline character.     "" "    def close (self, * Args, **kwargs): # real signature unknown          Close File         pass    def fileno (self,  *args, **kwargs): # real signature unknown          File Descriptor           pass    def  flush (Self, *args, **kwargs): # real signature unknown          flush File Internal buffers         pass     def isatty (Self, *args, **kwargs): # real signature unknown          determine if the file is consent to the TTY device         pass    def  read (Self, *args, **kwargs): # real signature unknown          read the specified byte data         pass     def readable (Self, *args, **kwargs): # real signature unknown          is readable         pass     def readline (Self, *args, **kwargs):  # real signature unknown          Read only one row of data         pass     def seek (Self, *args, **kwargs): # real signature  unknown         pointer position in specified file          pass   &nBsp;def seekable (Self, *args, **kwargs): # real signature unknown          whether the pointer is operable         pass     def tell (Self, *args, **kwargs):  # real signature unknown          Get pointer position         pass     def truncate (Self, *args, **kwargs):  # real signature  unknown         truncate data, preserving only the data that you specified before          pass    def writable (Self, *args, **kwargs): #  Whether real signature unknown         can write          pass    def write (Self, *args, **kwargs):  #  real signature unknown         Write Content         pass     def __getstate__ (Self, *args, **kwargs): # real  signature unknown        pass    def __ init__ (Self, *args, **kwargs): # real signature unknown         pass     @staticmethod  # known case of __ new__    def __new__ (*args, **kwargs): # real signature  unknown         "" " create and return a new  object.  see help (type)  for accurate signature.  "" "         pass    def __next__ (Self, *args, **kwargs):  # real signature unknown         "" " implement next (self) . " ""         PASS    DEF __REPR__ (Self, *args ,  **kwargs): # real signature unknown         "" " return repr (self) . " ""         pass     buffer = property (Lambda self: object (), lambda self, v:  None, lambda self: none)   # default    closed =  Property (Lambda self: object (),  lambda self, v: none, lambda self:  none)   # default    encoding = property (lambda self:  object (),  lambda self, v: none, lambda self: none)   #  Default   &nbsP;errors = property (Lambda self: object (), lambda self, v: none,  Lambda self: none)   # default    line_buffering =  Property (Lambda self: object (),  lambda self, v: none, lambda self:  none)   # default    name = property (lambda self:  Object (),  lambda self, v: none, lambda self: none)   #  Default    newlines = property (Lambda self: object (),  lambda  self, v: none, lambda self: none)   # default     _chunk_size = property (Lambda self: object (), lambda self, v:  None, lambda self: none)   # default    _finalizing =  property (lambda self:  object (),  lambda self, v: none, lambda self: none)   #  default3.x

Five. Modification of documents

File data is stored on the hard disk, so there is only coverage, there is no modification so to speak, we usually see the modified files, are simulated out of the effect, specifically, there are two ways to achieve:

Method One:

Mode one: The contents of the file stored in the hard disk are loaded into memory, can be modified in memory, and then overwritten by memory to the hard disk (word,vim,nodpad++ and other editors).

Import OS # calls the System module with open (' a.txt ') as Read_f,open ('. A.txt.swap ', ' W ') as Write_f:data=read_f.read () #全部读入内存, if the file is large, it will Very data=data.replace (' Alex ', ' SB ') #在内存中完成修改 write_f.write #一次性写入新文件os. Remove (' a.txt ') #删除原文件os. Rename ('. a.txt . Swap ', ' a.txt ') #将新建的文件重命名为原文件方法一

Method Two:

Mode two: The contents of the file stored on the hard disk are read into memory line by line, the modification is completed to write a new file, and finally overwrite the source file with a new file

Import Oswith Open (' a.txt ') as Read_f,open ('. A.txt.swap ', ' W ') as Write_f:for line in Read_f:line=line.replace ( ' Alex ', ' SB ') write_f.write (line) os.remove (' A.txt ') os.rename ('. A.txt.swap ', ' a.txt ')

File operations in Python