Hi Harpreet,
I'd recommend you please review option #3 in your first answer, if table SET and NUPI, complete row needs to be scanned *unless* a unique index (USI) is also defined, in which case the duplicate row check will be performed on the USI alone. A quite common practice in physical modeling is to use the Foreign Key as a NUPI for best joining with a major entity on its Primary Key, and the Primary Key as a USI, to enforce data integrity and the added benefit of low resources utilization for duplicates checking.
Hi abbonizio,
I believe there's no something as an additional uniqueness check overhead for option #1, the duplicate row check is just performed on the UPI
↧