Bash script manipulate output

[myrdos]

Hello, I have a bash script that runs two programs.

I take the output of the first program, and re-direct it to an ouput file:

ns net.tcl $currentCAC > output

Then I do a grep on the output file to get the relevant info:

grep KEYWORD output

This outputs a line to the console:

KEYWORD So that's a buffering time of 90.9121 ms

I want to get the '90' part of that line, and store it in a bash variable. How can I do this?

[ghostdog74]

assuming that the line where KEYWORD is found is always the same number of fields. The only changing part is the '90' part which is field 8

awk '/KEYWORD/{print $8 }' output

[drl]

Hi.

Modern versions of grep can do this kind of extraction:

#!/bin/sh

# @(#) s1       Demonstrate grep selection of only the matching text.

set -o nounset
echo " sh version: $BASH_VERSION"

echo " KEYWORD So that's a buffering time of 90.9121 ms" |
egrep -o '[0-9.]+'

exit 0

producing:

% ./s1
 sh version: 2.05b.0(1)-release
90.9121

cheers, drl