Shell Programming and Scripting

I have a file with 100s of lines of text. I want to perform an extraction of this line:

Code:

Info bpzs(pid=2652) using 1000 bits

I have not been able to extract it. Should I try expr match or is there another method? This line has data both in front of and in back of it. I do not have grep -o as an option on my system. Any ideas are greatly appreciated.

Code:

grep 'Info bpzs(pid=2652) using 1000 bits' file

Or, what's the complication here - which part of the string do you want to extract?

Just that one line that I printed. I did try the grep, and it returns words that are both to the left and to the right of that string instead of just that string alone.

---------- Post updated at 12:38 PM ---------- Previous update was at 11:51 AM ----------

I also tried to grep for the single "word" bpzs and what I got was this:

Code:

 echo $a |grep '\<bpzs\>'

it returned bpzs but also everything else on the line. Same with grep -w. I don't have grep -o. I really want it to return just the line:

Info bpzs(pid=2652) using 1000 bits
or even bpzs(pid=2652) would work

I have also tried:

Code:

echo `expr match "$a" '\([b][p][z][s] [0-9]\)'`

But this returns "req id processing Info bpzs("

What is the proper regex to match this string?

This may not be the most elegant way to accomplish want you specified, but at least, it works:

Code:

$ a="Info bpzs(pid=2652) using 1000 bits"
$ sed -n "/$a/ s:.*\($a\).*:\1:p" file
Info bpzs(pid=2652) using 1000 bits

Thanks for everyone's help on this. I was able to compose a perl command to get exactly what I needed:

Code:

|perl -ne 'print  /.*(\bInfo.bptm\(pid......).*/' | tr -d "(Info=bzspid"

Code:

sed -n 's/.*\(Info.*bits\).*/\1/p' filename

If the PID could change:

Code:

sed -n 's/.*\(Info.*pid=[0-9]*.*bits\).*/\1/p' filename