Archives of the TeradataForum
Message Posted: Fri, 15 Nov 2002 @ 17:29:44 GMT
| Subj: || || Re: Rowid and hash functions |
| From: || || Belle, Patrick A |
You're absolutely correct. The primary index is a physical distribution attribute and is not the primary key. Teradata has another
clause to identify the primary key and we shouldn't confuse them. For unique identification, you need the primary key and for distribution,
you need the primary index. The confusion is caused by the fact that they do often overlap.
Teradata Certified Master