Using variables with awk

dots · 07-07-2009, 12:15 PM

I need to use the ps -eaf command and awk to provide a report that will look like this:
USER = dbus
PID PROGRAM TTY
1538 dbus-daemon (?)

USER = rangel
PID PROGRAM TTY
9531 SCREEN (?)
13434 /bin/bash (pts/18)
13502 /bin/bash (pts/22)

USER = root
PID PROGRAM TTY
1 /sbin/init (?)
2 [kthreadd] (?)
3 [migration/0] (?)
I understand that I'm going to need a while loop and I think I can do the formatting, but I'm not sure how to get the user id stored as a variable. Am I even looking at going about this the right way? What is the best way to do this? Thanks for the help!

acid_kewpie · 07-07-2009, 12:51 PM

if you want to use awk you wouldn't use a while loop, as awk is for processing text streams, so naturally iterates over its inputs. Additionally using ps will sort by uid with the right options so it's actually really simple from then on, to just keep track of the current uid and print a new header each time the uid of the current record changes.

Code:

{
  if (userid != olduserid) {
    # print the headers out
  }
  # print the process data
  
  olduserid = userid
}

I'm a bit concerned that this is homework though, so i'm not going into any more detail at the moment. Should be fairly simple from here though.

dots · 07-07-2009, 02:35 PM

It is homework, I'm just stuck! I'm not understanding how to save the user id in a variable. I tried several different ways. Our book has an example and I tried it with what I'm doing and get an error. This is how I tried:

ps -eaf | sort -k1 | awk '\
$1~ { user = $1; print "USER = " user }'
echo $USER #just to make sure it was storing the user id

We have lab time in class tonight so it isn't a big deal. I just thought I could get it done and not have to go tonight.

dots · 07-07-2009, 02:38 PM

Sorry, the following line was from trying something different(it's no longer in my script):
echo $USER #just to make sure it was storing the user id

Tinkster · 07-07-2009, 03:00 PM

And what's the ~ supposed to be operating on?