Re: Parsing the spool file.....

On Oct 5, 3:08 pm, "akkha1..._at_gmail.com" <akkha1..._at_gmail.com> wrote:
> The configuration approach you mention is very sensible. Normally under
> development, we create table, grant/revoke and other amazing feats on
> ah doc basis for fast turnaround time. How do we implement and
> propagate changes in an effective and efficient manner to our
> production environment?
>
> For me, I would develop a script and run it on our test enviornment. If
> there is any error, fix it, restore the system and rerun until no error
> is obtained.This would not only verify the script but force one to test
> the backup and restore procedures as well. Nowadays, since most tools
> are going to GUI-based, developing script is not that easy. However, we
> would adopt the similar approach, ie, document every step with response
> and capture every damned screen shot and run the document several times
> on our testing environment.
>
> Then when we apply to the production, we would just follow the
> procedure developed in the test environment and have as little
> unexpected surprise as possible.
>
> I would like to know if there is better alternative than this "dump ass
> approach".

Why not have the shell script sqlplus statement execute a batch file with the SQL in it that the shell script expects to find in the same directory. It would be cleaner and avoids having to escape any UNIX meta-characters that might appear in the text such as when you need to reference a v$ or x$ view.

sqlplus /nolog <<EOF
connect / as sysdba
start the_real_work_sql
EOF HTH -- Mark D Powell -- Received on Thu Oct 05 2006 - 18:48:46 CDT