Query not using remote table primary key.

Question

Hello all,
My following query is not using the PK of remote database. vem_business_event_attribute is having a PK on ATTRIBUTE_TYPE_ID and business_event_id. The table is last analyzed on 04-dec-2010. It has around 5 milloin rows.
select /*+ DRIVING_SITE(ve) */
ve.ATTRIBUTE_TYPE_ID ,
ve.BUSINESS_EVENT_ID ,
ve.ATTRIBUTE_VALUE ,
ve.ATTRIBUTE_AS_NUMBER
from
vem_business_event_attribute@hello ve,
vem_attribute_types@hello va,
delta_insert_businessevent de
where
     ve.ATTRIBUTE_TYPE_ID = va.ATTRIBUTE_TYPE_ID
ve.business_event_id = de.business_event_id
     and de.business_event_id < 90000000;
Pls help.
Regards

Answer 1

If you need to move 150 million rows from one database to another, it is going to take some time, and a full scan would definitely be the most efficient sql way to do it. I would be surprised if 15 inserts of 10 million rows each would be faster than a single insert of 150 million rows.
Depending on your exact requirements, it might be faster to use datapump to export the rows from the remote database and import them on the local database, but you will still be moving those 150 million rows across the network. You might get some advantage from the export/ftp/import option if you compress the dump file before the ftp, and uncompress it on the target, but that would depend on the additional time required to compress and uncompress.
John