Hi all,
We have an application which deletes a large number of rows from a
table. It would be faster to simply insert the rows that we want to keep
into a second table, drop the original table and then rename the second
table to that of the one we have just dropped.
The only downside that I
$ORACLE_HOME/rdbms/admin/utlrp will recompile all invalid objects.
Could you partition the table so that you only need to drop a partition instead of
deleting rows?
Jay
[EMAIL PROTECTED] 04/15/02 08:23AM
Hi all,
We have an application which deletes a large number of rows from a
table. It
Can't you just do a variant of
SET HEAD OFF TERMOUT OFF ECHO OFF
select 'alter '||decode(object_type,'PACKAGE BODY',' PACKAGE
',object_type)||
' '||object_name||' compile '||
decode(object_type,'PACKAGE BODY','BODY ','')||';'
from user_objects
where object_type in ('PACKAGE',
Unfortunately we do not have the partitioning option available. I should
mention that when I say I want to delete a large number of rows, the number
of rows I want to keep is even larger. (Delete, say, 3 million rows, keep
7 million rows) so the option that some other have suggested of
