在Python中打开输入和输出文件

Question

I have the following code which is intended to remove specific lines of a file. When I run it, it prints the two filenames that live in the directory, then deletes all information in them. What am I doing wrong? I'm using Python 3.2 under Windows.

import os

files = [file for file in os.listdir() if file.split(".")[-1] == "txt"]

for file in files:
    print(file)
    input = open(file,"r")
    output = open(file,"w")

    for line in input:
        print(line)
        # if line is good, write it to output

    input.close()
    output.close()

Answer 1

open(file, 'w') wipes the file. To prevent that, open it in r+ mode (read+write/don't wipe), then read it all at once, filter the lines, and write them back out again. Something like

with open(file, "r+") as f:
    lines = f.readlines()              # read entire file into memory
    f.seek(0)                          # go back to the beginning of the file
    f.writelines(filter(good, lines))  # dump the filtered lines back
    f.truncate()                       # wipe the remains of the old file

I've assumed that good is a function telling whether a line should be kept.

Answer 2

If your file fits in memory, the easiest solution is to open the file for reading, read its contents to memory, close the file, open it for writing and write the filtered output back:

with open(file_name) as f:
    lines = list(f)
# filter lines
with open(file_name, "w") as f:      # This removes the file contents
    f.writelines(lines)

Since you are not intermangling read and write operations, the advanced file modes like "r+" are unnecessary here, and only compicate things.

If the file does not fit into memory, the usual approach is to write the output to a new, temporary file, and move it back to the original file name after processing is finished.

Answer 3

One way is to use the fileinput stdlib module. Then you don't have to worry about open/closing and file modes etc...

import fileinput
from contextlib import closing
import os

fnames = [fname for fname in os.listdir() if fname.split(".")[-1] == "txt"] # use splitext
with closing(fileinput.input(fnames, inplace=True)) as fin:
    for line in fin:
        # some condition
        if 'z' not in line: # your condition here
            print line, # suppress new line but adjust for py3 - print(line, eol='') ?

When using inplace=True - the fileinput redirects stdout to be to the file currently opened. A backup of the file with a default '.bak' extension is created which may come in useful if needed.

jon@minerva:~$ cat testtext.txt
one
two
three
four
five
six
seven
eight
nine
ten

After running the above with a condition of not line.startswith('t') :

jon@minerva:~$ cat testtext.txt
one
four
five
six
seven
eight
nine

Answer 4

You're deleting everything when you open the file to write to it. You can't have an open read and write to a file at the same time. Use open(file,"r+") instead, and then save all the lines to another variable before writing anything.

Answer 5

You should not open the same file for reading and writing at the same time.

"w" means create a empty for writing. If the file already exists, its data will be deleted.

So you can use a different file name for writing.