Special Design Considerations for the Teradata Data Warehouse Environment
An earlier topic in this chapter (“Avoiding the Data Mart‑Centric Approach” on page 24) observed that the most frequently seen advice for how to design the database that supports a data warehouse is antithetical to the overriding reason for having a warehouse: unrestrained access to enterprise-wide business data for any possible question that could possibly be conceived.
Are there any special database design considerations for building a Teradata data warehouse? The short answer is no, there are no special concerns, but certain aspects of commonly recommended data warehousing strategies should be evaluated for their veracity before you begin designing the database that supports your data warehouse. For these evaluations, see:
The Teradata parallel technology is optimized to perform tasks in a normalized environment that other relational DBMSs cannot match with a denormalized schema. Teradata is also optimized to perform tasks in a denormalized environment. Among the special performance advantages built into the Teradata system are the following: star join and other join optimizations, full table scan optimization, specially designed index types, a full complement of SQL aggregate and ordered analytical functions, and the most sophisticated parallel‑aware SQL query optimizer available.