Databases Reference
In-Depth Information
12
Time Variant
Database Definition
If this Time Variant Solution Design has taken you out of your comfort
zone, then please pause, breathe deeply, and get ready to go back to your
comfort zone. This database design is more about fitting Instance Keys and
indexes into an existing data warehouse design than it is about creating a
new data warehouse design. If you're waiting for this chapter to tell you
to throw everything away and prepare to start over, you're going to wait a
very long time. Rather than throw away everything you know about your
data warehouse, including all those little idiosyncrasies that only a handful
of people know about, this chapter will assume that you have maintained
your data warehouse to the best of your ability and that your data ware-
house is the best it can be. For that reason, obviously you're not trying to
solve a data model or ETL problem with a Time Variant design. You are
simply trying to integrate Time Variance into an existing data warehouse.
If, however, you're starting from the beginning, please be aware that
everything in this chapter comes after the source system analysis,
conceptual data model, logical data model, and the first physical data
model. Time variance should be introduced somewhere around the
second physical data model. All these analysis and design methods are
explained in
Building and Maintaining a Data Warehouse
. This discus-
sion of time variance in a data warehouse does not bypass or replace any
of the analysis and design methods that are key to the success of every
data warehouse. Rather, this discussion of time variance in a data ware-
house is in addition to the data warehouse analysis and design methods.
For that reason, this discussion of data warehousing assumes a shared
level of knowledge. If this is your first topic on data warehousing, please
have a copy of
Building and Maintaining a Data Warehouse
handy for
reference purposes.
191
