My first Publication Agile-Data-Warehouse-Design-eBook | 页面 247

8
H OW M ANY
Design Patterns for High Performance Fact Tables and Flexible Measures
How many times must a man look up…
— Bob Dylan, Blowin’ in the Wind
Everything that can be counted does not necessarily count;
everything that counts cannot necessarily be counted.
— Albert Einstein
In this chapter we describe how the three fact table patterns—transaction fact
tables, periodic snapshots, and accumulating snapshots—are implemented to
efficiently measure discrete, recurring and evolving business events. We particu-
larly focus on the agile design of accumulating snapshots, by describing how the
requirements for these powerful but complex fact tables can be visually modeled as
evolving events using event timelines, our final BEAM ✲ modelstorming tool. We
also describe the BEAM ✲ notation for capturing fact additivity and fully docu-
menting the limitations of semi-additive facts, such as balances. We conclude with
techniques for optimizing fact table performance and multi-fact table reporting by
concentrating on design patterns for aggregates and other derived fact tables that
accelerate and simplify BI queries
Point in time event measurement
This chapter
covers techniques
for incrementally
designing and
developing high-
performance fact
tables and flexible
measures
Chapter 8 Design
Periodic measurement Challenges
Evolving process measurement At a Glance
Modeling evolving event milestones and duration measures
Incremental development of complex fact tables
Flexible fact definition
Fact table performance
Correctly querying multiple fact tables at once
Cross-process analysis using simple BI tools
227