Shell Programming and Scripting

I have a file (OMlog0) that is quite large, I need to find the line "Automatic Recharge Audit Process Finished" and output that line and the time stamp that occurs two lines previous. The line that I was using is "sed -n '/Automatic Recharge Audit Process Finished/,/No errors/p' /sn/log/OM* > ARe.out", but it doesn't give me the time stamp just the "Finished" line and the "No Errors" line.


excerpt from OMlog0:

+++ MRMCMAS 2008-05-05 09:56:10 ASRT #360578 cc1 LEAD >
REPT SPA=EPPSA24D
Automatic Recharge Audit Process Finished Correctly!
Errors found: No errors
END OF REPORT #360578++-

+++ MRMCMAS 2008-05-05 09:56:12 ASRT #360579 cc0 ACT >
** REPT NETWORK ASSERT=200, SPA=EPAY24D
Network Message Error - Invalid value BILLID in is41css7!t_answer
Subscriber ID = 4165551212
Call ID = IS41_1890072211480
Call Instance ID = 594903
Scenario Location=is41css7!t_answer_received
Trigger type =is41c_trty_t_answer
END OF REPORT #360579++-

+++ MRMCMAS 2008-05-05 09:56:12 ASRT #360580 cc0 ACT >
** REPT NETWORK ASSERT=200, SPA=EPAY24D
Network Message Error - Invalid value BILLID in is41css7!t_disconnect
Subscriber ID = 4165551212
Call ID = IS41_1890072211480


Desired output:

+++ MRMCMAS 2008-05-05 09:56:10 ASRT #360578 cc1 LEAD >
REPT SPA=EPPSA24D
Automatic Recharge Audit Process Finished Correctly!
Errors found: No errors

You can use 'grep' to print out the line you're looking, the two lines above it, and the line after it:

grep "Automatic Recharge Audit Process Finished Correctly" -B 2 -A 1 file.txt

+++ MRMCMAS 2008-05-05 09:56:10 ASRT #360578 cc1 LEAD >
REPT SPA=EPPSA24D
Automatic Recharge Audit Process Finished Correctly!
Errors found: No errors

Hope this helps.

Thanks, but it doesn't seem to work for my system.

I am running Solaris 10 am when I do a man grep, the -A and -B options don't exist.

when I execute the command that you provided I get the following

grep: can't open -B
grep: can't open 2
grep: can't open -A
grep: can't open 1
grep: can't open file.txt

thanks,
Al

I don't know Solaris, but if your grep supports the "-p" (paragraph) option you could use that. As it seems from the example the parts of the output you are interested in are delimited by blank lines (which is what defines "paragraphs" in the sense of the "grep -p").

I hope this helps.

bakunin

If you do man on grep you will see that B and A are not supported operands

I really appreciate everyone's help, but it Solaris doesn't seem to like the -p option either. Is there maybe another command either than grep that I should try?

Do something like this:

awk '/Automatic Recharge Audit Process Finished Correctly/ { \
print pre2 ; print pre1 ;print ;getline ;print ;next}\
{ pre1=$0; pre2=pre1}' /sn/log/OM* > ARe.out