Insert a line into the middle of a text file in Python

Question

I want to insert a line into the middle of a text file in Python, so I tried

with open(erroredFilepath, 'r+t') as erroredFile:
    fileContents = erroredFile.read()

    if 'insert_here' in fileContents:
        insertString.join(fileContents.rsplit('insert_here'))
        erroredFile.truncate()
        erroredFile.write(insertString)

However, insertString got written at the end of the file. Why?

---

As an aside, I tried to simply things, by just using strings, instead of files.

'123456789'.join('qwertyuiop'.split('y'))

gives

'qwert123456789uiop'

what happened to the 'y' ?

Answer 1

If you want to write in the middle of the file use

fileinput module.

import fileinput
for line in fileinput.input("C:\\Users\\Administrator\\Desktop\\new.txt",inplace=True):
    print "something", #print("something", end ="") for python 3

remember whatever you print that will go in the file.So you have to read and print every line and modify whichever you want to replace.Also useprint "asd",...the,at the end is important as It will preventprintfrom putting a newline there.

Answer 2

Although OS-level details of files vary, in general, when you have a file open in r+ mode and do some read or write operation, the "current position" is left after the last read or write.

When you did:

fileContents = erroredFile.read()

the stream erroredFile was read to the end, so the current position is now "at the end".

The truncate function defaults to using the current position as the size to which to truncate. Assume the file is 100 bytes long, so that the current position "at the end" is byte 100. Then:

erroredFile.truncate()

means "make the file 100 bytes long"—which it already is.

The current position remains at the end of the file, so the subsequent write appends.

Presumably you wanted to seek back to the beginning of the file, and/or use truncate(0) (note that just truncate(0) will, at least on Unix-like systems, leave the seek position at the end of the file so that the next write leaves a hole where the original data used to be). You could also be slightly more clever: if you're inserting, just overwrite-and-extend in place (no truncate is required at all).

(Joel Hinz already answered the second question, I see.)

Answer 3

test.txt

a
b
c
d
e

1. Read into a list then overwrite

def match_then_insert(filename, match, content):
    lines = open(filename).read().splitlines()
    index = lines.index(match)
    lines.insert(index, content)
    open(filename, mode='w').write('\n'.join(lines))


match_then_insert('test.txt', match='c', content='123')

Result

a
b
123
c
d
e

2. FileInput

from fileinput import FileInput


def match_then_insert(filename, match, content):
    for line in FileInput(filename, inplace=True):
        if match in line:
            line = content + '\n' + line
        print(line, end='')  # Redirect to the original file


match_then_insert('test.txt', match='c', content='123')

3. seek

def match_then_insert(filename, match, content):
    with open(filename, mode='rb+') as f:
        while True:
            try:
                line = f.readline()
            except IndexError:
                break
            line_str = line.decode().splitlines()[0]
            if line_str == match:
                f.seek(-len(line), 1)
                rest = f.read()
                f.seek(-len(rest), 1)
                f.truncate()
                content = content + '\n'
                f.write(content.encode())
                f.write(rest)
                break


match_then_insert('test.txt', match='c', content='123')

Compare

Method	Time/s
Read into a list then overwrite	54.42
FileInput	121.59
seek	3.53

from timeit import timeit
from fileinput import FileInput


def init_txt():
    open('test.txt', mode='w').write('\n'.join(['a', 'b', 'c', 'd', 'e']))


def f1(filename='test.txt', match='c', content='123'):
    lines = open(filename).read().splitlines()
    index = lines.index(match)
    lines.insert(index, content)
    open(filename, mode='w').write('\n'.join(lines))


def f2(filename='test.txt', match='c', content='123'):
    for line in FileInput(filename, inplace=True):
        if match in line:
            line = content + '\n' + line
        print(line, end='')


def f3(filename='test.txt', match='c', content='123'):
    with open(filename, mode='rb+') as f:
        while True:
            try:
                line = f.readline()
            except IndexError:
                break
            line_str = line.decode().splitlines()[0]
            if line_str == match:
                f.seek(-len(line), 1)
                rest = f.read()
                f.seek(-len(rest), 1)
                f.truncate()
                content = content + '\n'
                f.write(content.encode())
                f.write(rest)
                break


init_txt()
print(timeit(f1, number=1000))
init_txt()
print(timeit(f2, number=1000))
init_txt()
print(timeit(f3, number=1000))

Answer 4

Not a Python answer but it may widen your horizon. Use sed:

$ cat input.txt 
foo
bar
baz
INSERT HERE
qux
quux

$ sed '/INSERT HERE/anew stuff' < input.txt
foo
bar
baz
INSERT HERE
new stuff
qux
quux

The command a will append the text on a new line. If you want to insert the text before the match, use the command i:

$ sed '/INSERT HERE/inew stuff' < input.txt
foo
bar
baz
new stuff
INSERT HERE
qux
quux

Answer 5

Why do you not try a two steps solution? At first, you read and fix the string, at the second step, you rewrite the file. Probably it's not the most efficient algorithm, but I think it works.

with open(erroredFilepath, 'r') as erroredFile:
    fileContents = erroredFile.read()   
    fileContents.replace('insert_here', 'insert_string')

with open(erroredFilePath, 'w') as fixingFile:
    fixingFile.write(fileContents)