CODEWITHSUNDEEP
pythonsubprocessstderr
Richard
Richard

Reputation: 65530

Python: only print stderr, not stdout, from subprocess.call()?

I'm using Python's subprocess.call (docs here) and I have a question about standard output.

I have the following command: 

subprocess.call("wget http://google.com", shell=True)

It works great, but it prints the output of wget, so I get an absolute ton of wget output like: 

--2013-06-28 08:54:47--  http://google.com/
Resolving google.com... 173.194.41.100, 173.194.41.98, 173.194.41.97, ...
Connecting to google.com|173.194.41.100|:80... connected.
HTTP request sent, awaiting response... 301 Moved Permanently
.... etc

How can I change the call so it only prints anything if there is an error?

I have tried studying the subprocess docs, but I think I'm too inexperienced to figure it out for myself.

Upvotes: 0

Views: 1082

Answers (2)

Joachim Isaksson
Joachim Isaksson

Reputation: 180917

wget normally reports all output on stderr, so redirecting stdout will not help with getting errors only.

You can either use -nv (non verbose) or -q (quiet) as options to wget, the difference being that -q does not even report errors while -nv reports errors or a single success line to stderr.

The return code is unchanged in both cases, so it can be used to detect any errors if that's the way you prefer.

EDIT: If you really just want to ignore all output from wget, you can redirect stderr and stdout to a null device;

import subprocess
import os

DEVNULL = open(os.devnull, 'wb')

subprocess.call("wget http://google.com/", stdout=DEVNULL, stderr=DEVNULL, shell=True)

In my Python 3.3.1 install, DEVNULL can be imported from subprocess instead, but I can't find documentation on exactly what version it was added in. The above works in both Python2 and Python3.

Upvotes: 1

lyager
lyager

Reputation: 26

Actually I think the easiest way would be to use the '-q' option that 'wget' has. You can still catch error by looking at the return value.

Upvotes: 1

Related Questions