Depending on your setup, Flashback Query may be of some benefit:
http://download.oracle.com/docs/cd/B10501_01/server.920/a96524/c21cnsis.htm#20178
Of course, at the risk of sounding like I'm shutting the gate after
the horse has bolted, your main concern is how this happened and how
you are going to prevent it happening again. You may want ot examine
your backup solution, too ;)
HTH
-g
Since this does not sound like a real production system (scott/tiger!!!) I
hope you have a backup!
Shakespeare
a.. Flashback query does not work through DDL operations that modify
columns, or drop or truncate tables.
I fear the worst....
Shakespeare
You are, of course, correct. I hope he's got a backup :-)
-g
Yes restoring the backup is the solution.
It is good practice as well to do a restore regularly to make sure you
know how to do it. (so when you need to do it in production, you are
confident of your restore process when the pressure is on!)
Ed
anujit, Shakespeare brings up the first thought I had which is why is
the scott/tiger username and password being used. This is a training
Id for which the password should always be changed in test and the ID
should be dropped or at least locked out of production.
On a production system you should have rman or at least hot backs that
could be used to restore the database to a duplicate then advanced to
a point just before the objects were dropped.
Do you take exports? For test this should be good enough.
This is all providing you actually determined that you really need to
restore these objects.
HTH -- Mark D Powell --