Abstract
The most common implementation platform for multidimensional data warehouses is RDBMSs storing data in relational star and snowflake schemas. DW schemas evolve over time, which may invalidate existing analysis queries used for reporting purposes. However, the evolution properties of star and snowflake schemas have not previously been investigated systematically. This paper systematically investigates the evolution properties of star and snowflake schemas. Eight evolution operations are considered, covering insertion and deletion of dimensions, levels, dimension attributes, and measure attributes. For each operation, the formal semantics of the changes for star and snowflake schemas are given, and instance adaption and impact on existing queries are described. Finally, we compare the evolution properties of star and snowflake schemas, concluding that the star schema is considerably more robust towards schema changes than the snowflake schema.
Bidragets oversatte titel | Skemaevolution for stjerner og snefug |
---|---|
Originalsprog | Engelsk |
Titel | Proceedings of the Sixth International Conference on Enterprise Information Systems |
Antal sider | 9 |
Publikationsdato | 2004 |
Sider | 425-433 |
Status | Udgivet - 2004 |
Begivenhed | ICEIS 2004 - Porto, Portugal Varighed: 14 apr. 2004 → 17 apr. 2004 Konferencens nummer: 6 |
Konference
Konference | ICEIS 2004 |
---|---|
Nummer | 6 |
Land/Område | Portugal |
By | Porto |
Periode | 14/04/2004 → 17/04/2004 |