Hugo and Hakan, both bring up good points. There may be justifiable cases to not have a clustered index on a table - in an ideal world where one has thought long and hard about the decision. Besides the ability to rebuild indexes online, or defrag the indexes in the table; we also have the added benefit of being able to more easily move tables to new filegroups - should we desire or the need arise. Too often it is a novice at database design creating new tables or even databases on the whole, and oftentimes all indexes are overlooked as well as keys of any sort (PK or FK). And agreed, it should be evaluated by the team (dev and dba) whether to use a clustered index in place of the non-clustered index - just in case there was a valid reason to just use non-clustered indexing.:hehe: