Run awk file from shell script without specifying awk's exact location

Question

I'm trying to debug a shell script that calls an awk file. Which is a nightmare, cause I've never used either before, nor am I very fluent with linux, but anyway

A dev made an awk file and is trying to run it in a shell script.

To try and run it from a separate location, without needing to specify the exact location, they put the awk script in a folder that's in the PATH variable. So that awk file should be available everywhere, right?

When you run it like this...

awk -f script.awk arg1

...can awk find that script? It spits out an error, when the shell script tries to run that awk command:

awk: fatal: can't open source file `script.awk' for reading (No such file or directory)

Answer 1

As you know, awk can't find the script itself.

If the script is marked as executable, and if you have a which command, then you should be able to do:

awk -f `which script.awk` arg1

Alternatively, and probably better, make the script into an executable:

#!/usr/bin/awk -f 
BEGIN { x = 23 }
      { x += 2 }
END   { print x }

The shebang needs the '-f' option to work on MacOS X 10.7.1 where I tested it:

$ script.awk script.awk
31
$

That gives you a self-contained single-file solution.

Answer 2

No, that's not going to work. awk needs the path to the script to run, it won't use the PATH variable to find it. PATH is only used to find executables.