How to capture the output from "subprocess.call" to a file?

Question

In my code I have a line similar to this:

rval = subprocess.call(["mkdir",directoryName], shell=True)

and I can check rval to see if it is 0 or 1, but if it is 1, I would like to have the text from the command "A subdirectory or file ben already exists." in a file format, so I can compare it to another file if I want to make sure the text is the same.

Is it possible to have a line like this, but I know this does not work

rval = subprocess.call(["mkdir",directoryName], shell=True) >> filename

so no matter what happens with the command, the text is captured in filename, and rval still has the return code?

Duplicate of all of these: http://stackoverflow.com/search?q=%5Bpython%5D+subprocess+capture — S.Lott, Oct 20 '10 at 16:04
possible duplicate of [Capture subprocess output](http://stackoverflow.com/questions/2525263/capture-subprocess-output) — S.Lott, Oct 20 '10 at 16:04
I am doing this for a command I can not post here, but it has similar characteristics as mkdir. — Dag, Oct 20 '10 at 16:16
@S.Lott, it's *almost* a duplicate. The difference would be `Popen` vs `call` (which I agree is subtle), and the desire to write to a file rather than capturing the output within the program. Even though it seems the questioner would rather capture the output, it's useful to answer the question as-is. — Mark Ransom, Oct 20 '10 at 16:28

score 17 · Answer 1 · answered Oct 17 '12 at 05:18

17

The subprocess module has a built in 'check_output' function for doing this:

In [11]: result = subprocess.check_output(['pwd'])

In [12]: print result
/home/vagrant

answered Oct 17 '12 at 05:18

btubbs

1,845
1
16
19

1

But then you lose the exit value :( – speg Apr 25 '13 at 18:10
3

Note that check_output is only possible since 2.7 – Doomsday Sep 04 '13 at 09:51

score 15 · Accepted Answer · answered Oct 20 '10 at 16:23

15

import subprocess
f = open(r'c:\temp\temp.txt','w')
subprocess.call(['dir', r'c:\temp'], shell=True, stdout=f)
f.close()

answered Oct 20 '10 at 16:23

Mark Ransom

299,747
42
398
622

1

Any reason why you used 'shell=True' in this case? Cause I read using that is a bad idea (on *nix at least). – user225312 Oct 20 '10 at 17:41
@PulpFiction, two reasons: first it was included in the original question, second I was using a Windows shell command for testing. – Mark Ransom Oct 20 '10 at 17:42

Jatin Kumar · Answer 3 · 2015-07-17T00:08:05.147

12

import subprocess

try:
    result = subprocess.check_output(['dir', r'c:\temp'], shell=True)
    print result
except subprocess.CalledProcessError as e:
    return_code = e.returncode

You anyway need to use try catch because it throws exception if return code is non zero :)

edited Jul 17 '15 at 00:08

answered Sep 09 '14 at 07:05

Jatin Kumar

2,635
9
36
46

1

It would be simpler to just use `except CalledProcessError as e: returncode=e.returncode`. – orodbhen Jul 16 '15 at 13:58

How to capture the output from "subprocess.call" to a file?

3 Answers3

Linked