By: dataismypassion

dataismypassion — Fri, 13 Apr 2012 16:08:53 +0000

I have to believe the column(s) that will be target of your primary key has data that is unique. If that is the case can you live with creating a non-clustered unique index on the column(s) ? If one of the purposes of wanting to create a primary key in the first place is to ensure row uniqueness then you will have satisfied your condition. Even if your reason is to improve performance, you will have created an index that will be the basis for retrieving data rather than retrieving from the heap.
You may also want to check if your tempdb is constrained on max size. Index creation sometimes use tempdb.

Comments on: Creating clustered index on exiting table

By: dataismypassion