Read lines from file and subsequently only write those lines to the file which don't contain a certain phrase

Question

I have one column list file:

$ cat test
blah
blahblah
blah22
bluh
blih
blihihih

All I want to is to remove blah-like lines. (Equivalent of bash grep -v blah.) So that blah, blahblah and blah22 are removed.

I have this so far:

>>> f = open('test', 'rw+')
>>> line = f.readlines()
>>> for l in line:
...     if l.find('blah') == -1:
...             f.truncate(0)
...             f.write(l)
... 
>>> f.close()  

I thought it would be fine, but after running this, from test file only this line is left:

$cat test
blihihih  

How come, this has removed blih or bluh?

Answer 1

How come, this has removed blih or bluh?

To answer your question, you truncate the file too often. Every call to f.truncate(0) will reset the file size to 0. Therefore only the last line blihihih will survive.

If you want to do it in one go:

f = open('test.txt', 'r+')
lines = f.readlines()
f.seek(0)
for line in lines:
    if line.find('blah') < 0:
        f.write(line)
f.truncate()
f.close()

Additionally, you should really use with:

with open('test.txt', 'r+') as f:
    lines = f.readlines()
    f.seek(0)
    for line in lines:
        if line.find('blah') < 0:
            f.write(line)
    f.truncate()

(You won't need to write f.close() then. Read more about why using with here for example.)

Answer 2

I think is best if you open the file first, and then write to it:

# open file
lines = open('test').readlines()

# over-write the file
with open('test', 'w') as out:
    for line in lines:
        if 'blah' not in line:
            out.write(line)

Output (in test)

bluh
blih
blihihih