Question

I have a text source (a log file), which gets new lines appended to it by some third party.

I can output the additions to my source file using tail -f source. I can then pipe that through an awk script awk -f parser.awk to parse and format the output.

My question is: while tail -f source | awk -f parser.awk is running, is there a way to call function foo() inside my parser.awk script every time there is more than 5 seconds elapsed without anything coming through the pipe into the standard input of the awk script?

Edit: Currently using GNU Awk 3.1.6. May be able to upgrade to newer version if required.

Answer 1

If your shell’s read supports -t and -u, here’s an ugly hack:

{ echo hello; sleep 6; echo world; } | awk 'BEGIN{ 
    while( "while read -t 5 -u 3 line; do echo \"$line\"; done" | getline > 0 )
        print
    }' 3<&0

You can replace the print in the body of the while loop with your script. However, it would probably make a lot more sense to put the read timeout between tail and awk in the pipeline, and it would make even more sense to re-implement tail to timeout.