Thanks! That's definitely worth a try. I'll add this to our large refreshes to see what difference it makes tonight. - Maureen Allen, Brandon wrote:
The solution/workaround is to set atomic_refresh=>false so it will truncate the mview table instead of doing a delete. Just search for "atomic_refresh" on MOS and you'll see the relevant notes about this (e.g. 365157.1). Regards, Brandon
-- //www.freelists.org/webpage/oracle-l