Well, a few points:

- Bulk inserts are faster than single inserts. Of course, the problem is that you may blow your logs if you try to insert 1 million records, I don't know. See if you can have them make you a special rollback segment (or whatever) for this statement, because the bulk insert will be *much* faster.

- NOT IN's are generally not very efficient. The outer-join trick should be used instead.

Here is a statement that incorporates both suggestions:

-INSERT INTO
- WEBLOG
- (
- URL ,
- UNIVERSITY ,
- SERVER_NAME ,
- SEQNUM
- )
- SELECT
- W1.URL ,
- W1.UNIVERSITY ,
- W1.SERVER_NAME ,
- W1.SEQNUM
- FROM
- WEBLOG@STATS W1,
- WEBLOG W2
- WHERE
- W2.SEQNUM (+)= W1.SEQNUM AND
- W2.SEQNUM IS NULL


HTH,

- Chris