For the design of data structures of a data warehouse, star scheme techniques are becoming more important every day. This course treats dimensional modeling in detail, including the latest developments (e.g. data vault of Dan Lindstedt).
Designers of data warehouses.
Basic understanding of relational database systems is required (see SQL and relational databases fundamentals).
- Star schemes:
What is dimensional modeling? • Why a star scheme? • relational versus multidimensional databases • facts table, dimension tables • dimensions: causal, heterogeneous, monster, mini, degenerated • normalisation, denormalisation, snowflake schemes • pure star scheme versus snapshot.
- History in a data warehouse:
Detail versus analytic attributes • mutations in dimensions, type 0 to 5 approach • minidimensions and history • transaction orientation versus state orientation • scheme changes through time
- The OASI concept:
The One Attribute Set Interface • technical versus user dimensions • join navigation, aggregate navigation • OASI and heterogeneous dimensions
- Design, starfinding:
DWH-Blueprint, evolutionary development • conform dimensions • sample - snapshot - adapting a star scheme • OLAP design • approach for dimensional modeling • starfinding
- Advanced concepts:
Bridge tables • complex hierarchies • star schemes and meta-data • extraction, cleansing, transformation and loading
- User disclosure:
Middleware • protection against 'runaway' queries • OLAP tools • report tools
The theory will be illustrated by exercises. As the number of attendees is limited, all students will be individually coached during these exercises. Furthermore, an extra instructor will be present when needed. In this way, everyone will get optimal benefit from this course, even if there are differences in the background of the students.
Harm van der Lek.
|SESSION INFO AND ENROLMENT|