Archives of the TeradataForum
Message Posted: Thu, 01 May 2003 @ 17:30:54 GMT
| Subj: || || Re: Surrogate Keys vs Natural Keys |
| From: || || Judge, James A |
I agree with both points. I think the issue arises in the physical schema design. As you know a Teradata table's "primary index" is a
key PDBD consideration. Co-location of rows among tables that are frequently joined using the PK-FK associations is "usually" implemented
with the goal of having the joined tables with the same "primary index" column(s) whenever possible. In R4 and even more so in R5
"alternate" indexes (join, value-ordered, partial covering, etc.) give you more options to support the PK-FK associations for PDBD.
If your model (as most do) have a unique surrogate key for each entity don't just default to making the surrogate key the primary index
in every entity.