Thread: awk on non stop stream

Hello,

I run an application which output is non-stop. Is it possible to use awk or something else to see only selected columns?

thx

well i use

Code:

 application | cut -f1 | awk {'print $1'}

but it doesn't work. Why?

If u let me know what your trying to monitor with the awk script i can write u a one line perl script ??

Newk

well the ouput have 11 columns. I want only the 1,2,6,8,11.

ok, as far as i can see from a few tests

what u need is

application | awk '{ print $1, $2, $6, $8, $11}'

that should work

Newk

I tryed already, it doesn't work :/ maybe (?) because the applications output is non stop?

Do u need the formatted output as it comes out?? or wud after its run work too?

Newky

I need the formatted output as it comes out.. :/

I'm forgetting my shell scripting, but what about something like

Code:

application& > temp_file
while true; do tail temp_file | awk whatever; done

can i do it without creating a file? is it possible something like that application | awk ... ?

Pipe isn't going to pass anything along until the first process terminates, I'm pretty sure.

dawm :/ i don't want to create a file because the stream is non-stop and if the application runs for a long time it'll create a huge file :/. There isn't something like the line-line buffer you suggested me before?

You probably then want to redirect the output of your application to your script. Maybe something like

Code:

while true; do
    read var1 var2 var3 var4 var5 var6 var7 var8 var9 var10 var11; #There's probably a shortcut for this
    echo "$var1 $var6 $var8 $var11";
done

the problem is that i call the application using popen inside a c program. Can i use something like that?