Question

I’m a python novice and am struggling with some concepts here – any help is appreciated.

I’ve got a custom system tool that queries a database, and returns several lines as results to be read — one on each line. The following python script accepts the site FQDN from raw_input and runs $path on that fqdn.

#!/usr/bin/python

import subprocess
import getpass

#get the site name.
site = raw_input("What is the name of the site?: ").strip()

#run path.
cmd = 'path '+ site;
p = subprocess.Popen(cmd, shell=True, stdout=subprocess.PIPE);
path_output = p.stdout.read().strip().split('\n')

print path_output

Which returns results like this this:

['    fqdn          = www.hcasc.info', '    account_id    = 525925', '    parent_id     = 525925', '    nfs           = /mnt/stor7-wc2-dfw1/525925/www.hcasc.info', '  server_type   = PHP5', '    ssl           = False', '    host_ip       = 98.129.229.186', '    cgi_hosting   = False', '    test_link_ip  = 98.129.229.186', '    ipv6_ip       = 2001:4800:7b02:100::1600:0']

How can I get that extra whitespace out from the “nfs = etc”, or just take the third column (aka awk ‘{print $3}’) and/or assign each piece of these results from bash to separate variables for further manipulation?

Just having some trouble mounting this learning curve, your help is sincerely appreciated.

Answer 1

The third column would be line.split()[2]; if you want to throw out the first two words and take the rest, it’s line.split(None, 2)[-1]. (split with no arguments, or a None as the first argument, splits on any whitespace.)

>>> '    fqdn          = www.hcasc.info'.split()
['fqdn', '=', 'www.hcasc.info']

>>> for var, equals, rest in (l.split(None, 2) for l in path_output):
    assert equals == '='
    print var, 'is', rest

fqdn is www.hcasc.info
account_id is 525925
parent_id is 525925
nfs is /mnt/stor7-wc2-dfw1/525925/www.hcasc.info
server_type is PHP5
ssl is False
host_ip is 98.129.229.186
cgi_hosting is False
test_link_ip is 98.129.229.186
ipv6_ip is 2001:4800:7b02:100::1600:0

---

Explanation:
(l.split(None, 2) for l in path_output) is a generator expression, which runs l.split(None, 2) for each value of path_output (calling it l). It’s like a list comprehension, which is the same thing but with [] around it instead of (), but it only runs the l.split call as the for loop passes over it and then forgets the previous values, whereas the list comprehension would construct one big list with all the results of l.split at each step first, and then loop over that list normally. This way is like doing

for line in path_output:
    var, equals, rest = line.split(None, 2)
    ...

but a little shorter. 🙂

---

If you want to put this into a dictionary, as DSM suggests, you could do this in this way (just for context) as

d = dict((var, rest) for var, equals, rest in (l.split(None, 2) for l in path_output))

or, in Python 2.7 / 3, the much nicer

d = { var: rest for var, equals, rest in (l.split(None, 2) for l in path_output) }

Of course, you could make this a little more readable on two lines:

output_vals = (l.split(None, 2) for l in path_output)
d = dict((var, rest) for var, equals, rest in output_vals)

Whether you want a dictionary or just a loop depends on what processing you’re going to be doing with it, but the dictionary is probably a better approach for most things.