Trouble storing a variable in a shell script after awk output

uncle-c · 02-08-2010, 06:03 AM

Hi folks,

On command line I have no problem storing a variable e.g

Code:

:~/bin$ process=`ps -ef | awk ' $8 == "idesk" { print $2 }'`
:~/bin$ echo $process
26736
:~/bin$

When I try to incorporate this into a shell script I get a blank.

Code:

 :~/bin$ cat process_idesk
#!/bin/bash

process=`ps -ef | awk ' $8 == "idesk" { print $2 }'`

Code:

 :~/bin$  ./process_idesk
:~/bin$ echo $process

:~/bin$

The process_idesk script has been chmoded to be executable by the user. I'm sure there must be a silly omission on my behalf.

Thanks

C

druuna · 02-08-2010, 06:22 AM

Hi,

The script runs in its own (child) shell. After the script is finished, you return to your initial (mother) shell.

The process variable is filled in the child shell and will not be available in the mother shell after execution.

Put the echo $process in the script and it will echo what you expect. If you need the use the variable outside the script you need to store it first (to file) to be able to use it afterwards.

Maybe this will help: Advanced Bash-Scripting Guide - Chapter 20. Subshells

Hope this clears things up.

uncle-c · 02-08-2010, 06:35 AM

Ah I see ! That explains it. It makes sense now. Thanks druuna !

C

catkin · 02-08-2010, 07:13 AM

One solution is to source the script -- then all the variables set in the script are defined in the parent shell.

Code:

source ./process_idesk