[File Input Module]Replacing string in a file

[Steven D'Aprano]

On Thu, 28 Jan 2010 21:36:27 -0800, vanam wrote:

> Hi all,
> 
> For replacing string in the file, i have used the module File input for
> replacing the string in the file.

If you're processing a single file, using fileinput is overkill, and slow 
as well. Here is a quick-and-dirty way of processing one file, with no 
error checking:

text = open("some file", "r").read()
# modify the text however you like
text = text.replace("ham", "spam")  
# replace in place
open("some file", "w").write(text)


In a real application, you would need to do significant error checking. 
But it shows the basic technique: you read a file, modify it in memory, 
then write it back to disk.


> For understanding and execution purpose, i have just included Python as
> a string in the file and want it to be replaced to PYTHON.
> 
> Below are my queries and code: (Correct me if my understanding is
> wrong???????)
> 
> 1))
> 
> import fileinput
> x = fileinput.input('data.txt',inplace=0)
> for line in x:
>     line = line.replace('Python','PYTHON)
>     print line,
> x.close()
> 
> The above piece of code will not create any backup file but  it will
> replace PYTHON (Print on the console) but not physically write to the
> file.

Correct.


> 2)))
> 
> import fileinput
> x = fileinput.input('data.txt',inplace=1) 
> for line in x:
>    line = line.replace('Python','PYTHON') 
>    print line,
> x.close()
> 
> The above piece of code will create backup file but hidden (in the form
> of bak file) and it will physically write to the file -- I have verified
> the contents of data.txt after the file operation and it had written
> successfully.But why it is not printing line i.e. string in the file on
> the console.

You should read the documentation to fileinput.

The inplace=1 argument tells fileinput to redirect stdout to the current 
file, so the output of print will go to the file instead of the console.

Personally, I find that an incredibly unintuitive and dangerous API, and 
would recommend against fileinput just for that reason alone.



> 3)))
> 
> import fileinput
> x = fileinput.input('data.txt',inplace=1) 
> for line in x:
>    line = line.replace('Python','PYTHON')
> x.close()
> 
> The above piece of code after execution is wiping out the full contents.
> But addition of print line, is exactly replacing the string, what
> exactly addition of print is making difference???

That's how the module works: printing will go to the file instead of the 
console. Since you don't print anything, the file gets written with the 
empty string.


> 4)))
> 
> import fileinput
> x = fileinput.input('data.txt',inplace=1,backup='content.txt')
> for line in x:
>    line = line.replace('Python','PYTHON') print line,
> x.close()
> 
> The above piece is creating a backup file by name data.txtcontent.txt 

Yes, that's write, you told it to use "content.txt" as the file extension 
on the backup file. So it starts with the existing name, "data.txt" and 
adds "content.txt" as the extension, giving "data.txtcontent.txt".

You really should read the documentation. In the interactive interpreter:


import fileinput
help(fileinput)


> (I am not sure whether created file name is correct or not?) and to the
> back up file it had added previous content i.e., Python and it had
> replaced the contents in data.txt to PYTHON

Is there a question there?



-- 
Steven