Shell Programming and Scripting

I am spooling a file from oracle sqlplus, but I need to grep out certain things. The problem is that if i find a keyword i need to get several lines worth of stuff and redirect it to a file.

I am not sure how to do this.

Here is sample output.

Code:

Total missing in LDAP: 0
PL/SQL procedure successfully completed.

BILL
BOB
TED
JAMESS
Total records: 4
PL/SQL procedure successfully completed.
 
ROB
ALICE
Total records: 2
PL/SQL procedure successfully completed.
Total records: 0
PL/SQL procedure successfully completed.

Any time total records is > 0, I need the total number of records line and I need each of the names. There will be 1 name per line and the total number of names will be equal to the total records.
There will be a line that says

PL/SQL procedure successfully completed.

between each Total Records output.

is this an egrep thing? or just grep? I am not sure how to do this.

There is an egrep you can use. Try this:

Code:

egrep -v "successfully completed|Total records: 0" my_logfile > new_file

This will remove the lines you didn't need and direct them to 'new_file'.

As a note: do you know about the utl_file package? It allows you to write directly to a file, so that BOB CAROL TED and ALICE will appear in a separate file, saving you this kind of post-processing.

utl_file is not a good option. i need to run this in alot of databases. this is why I want to do it in the shell. so I would need to make sure I have appropriate directories set up in lots of databases and this requires written approval in production.

this way I can just do it.

thanks for the egrep. worked great guys!