在python 3.3中将字符串写入文件失败

Question

I'm using this python code to extract email addresses from a given file (source_file.txt) and to write those email addresses in a separate file. (I'm using python 3.3)

import urllib.request
import re
import fileinput, glob, string, sys, os
from os.path import join
import os

filePath = "source_file.txt"

if not filePath:
    print("Sorry! Source File could not be located!")
else:
    page = open(filePath, "r")
    pageContent = page.read()
    page.close()

style_normal = re.compile("[-a-zA-Z0-9._]+@[-a-zA-Z0-9_]+.[a-zA-Z0-9_.]+")
style_text_one = re.compile('[-a-zA-Z0-9._]+\s+at\s+[-a-zA-Z0-9_]+\s+dot\s+[a-zA-Z0-9_.]+')
style_text_two = re.compile('[-a-zA-Z0-9._]+\(at\)[-a-zA-Z0-9_]+\(dot\)[a-zA-Z0-9_.]+')

style_normal_list = style_normal.findall(str(pageContent))
style_text_one_list = style_text_one.findall(str(pageContent))
style_text_two_list = style_text_two.findall(str(pageContent))

f = open('emails_file.txt', 'a')
f.write('testing')

for item in style_normal_list:
    print("%s" %item)
    f.write("%s" %item)
for item in style_text_one_list:
    text_one = item.replace(' at ','@')
    text_two = text_one.replace(' dot ','.')
    print(text_two)
    f.write(text_two)

for item in style_text_two_list:
    text_one = item.replace('(at)','@')
    text_two = text_one.replace('(dot)','.')
    f.write(text_two)

The problem is that when I run this the file gets created, but nothing gets written in the file. Not even the text 'testing'. File is created but it is blank.

I have used the print statements to check whether other parts of the code is working correctly. According to the python doc the syntax is correct. Can any of you point out any mistakes I have made in this code?

Answer 1

When you call write on a file, the data doesn't actually get written to disk immediately, it gets stored in a buffer. Every so often, the buffer is flushed. You can explicitly call flush if you want to control it, but it's usually better not to—when you close the file, it automatically flushes.

But you have to close the file for this to work. When a file object gets deleted, it closes itself, but you can't rely on the file object getting deleted at any particular time, or even ever. (You can learn the details of how it works in CPython 2.7, CPython 3.3, IronPython, etc., and how CPython 3.3 handles exit differently on Unix vs. Windows, and so on, but you still wouldn't want to rely on those details.)

So, if you add an explicit f.close() at the end of your code, that almost solves the problem… but not completely. What if some other line raises an exception? Then it will never get to the close() call.

You can solve this with try / finally , but there's a better way, the with statement:

with open('emails_file.txt', 'a') as f:    
    f.write('testing')
    # ...
    for item in style_text_two_list:
        text_one = item.replace('(at)','@')
        text_two = text_one.replace('(dot)','.')
        f.write(text_two)

This guarantees that as soon as you get to the end of the indented with block—even if that happens because of an exception, or quitting the program (or return ing from the function, if you're in a function, or break ing from the loop, if you're in a loop, etc.), f will be closed.

This is all explained in Methods of File Object in the official tutorial, but looking over how it's explained, I can imagine a lot of new programmers wouldn't get any of that out of what's written. And the with statement documentation only really makes sense if you already understand what it does, and likewise for the explanations about file objects being buffered… PEP 343 , which introduced the with statement is a little better, but it spends as much time talking about what's wrong with the two previous related PEPs as about why the feature is useful. So, I guess it's understandable that many developers don't know any of this.