The document discusses using aggregates in a data warehouse for improved query performance. It describes requirements for an aggregate navigator tool that would transparently determine the best aggregate table to use for a given query based on available dimensions and facts, table sizes, and availability of aggregates. This allows queries to always use the most performant aggregate while abstracting aggregate selection complexity from users and applications.
2. Using Aggregates
the proper use of aggregates introduces a new set of
requirements for the data warehouse architecture.
Mechanisms must be in place to ensure that the most
appropriate table is used to respond to every warehouse
query.
2
4. The Schema Design
Dimensionally designed aggregates can answer a question
if all facts and dimensions required by the question are
present in the aggregate.
Order dollars by month for 2006
4
5. Relative Size
Use the relative number of rows in each aggregate fact
table or pre-joined aggregate to estimate the
performance ranking of each option.
Order dollars by month for 2006
5
6. Aggregate Portfolio and Availability
Aggregates should be chosen each time a query is run.
This allows consideration of availability status and
accommodates a changing portfolio of aggregates.
6
8. aggregate navigator
Evaluating schema design, table size, and availability can be
confusing for an end user. And since these evaluations
should be made each time a query is run, how is a query
to be saved in a report? The aggregate navigator solves
these problems.
8
9. REQUIREMENTS FOR THE AGGREGATE
NAVIGATOR
The aggregate navigator should permit users and
applications to deal with the base schema only.
The aggregate navigator should support aggregate star
schemas (including aggregate dimensions) and pre-joined
aggregates. It should not require changes to the base
schema design.
The aggregate navigator should be able to identify all
necessary information automatically, through examination
of the database catalog. This includes aggregate families,
conformance, and relative table size.
9
10. REQUIREMENTS FOR THE AGGREGATE
NAVIGATOR
The aggregate navigator translates base-schema SQL into
aggregate-aware SQL at runtime. This process is
transparent to the applications that issue SQL queries.
The aggregate navigator should respond to the dynamic
availability of aggregates.
The aggregate navigator should facilitate easy addition and
removal of aggregates from the data warehouse.
A single aggregate navigator should service all front-end
applications.
A single aggregate navigation system should service all
back-end databases.
10
11. Bibliograf鱈a
Mastering Data Warehouse Aggregates.Solutions for Star
Schema Performance. Christopher Adamson.
11