Chapter 5 Temporal Data Warehouses Copyright 2008 Elzbieta
Chapter 5 Temporal Data Warehouses Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 1
(a) Type 1 (b) Type 2 (c) Type 3 Fig. 5. 1. Three different implementation types of slowly changing dimensions Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 2
(a) Events (b) States Fig. 5. 2. Representing the temporal characteristics of real-world phenomena as events or as states Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 3
Fig. 5. 3. Temporal data types Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 4
meets overlaps/intersects contains/inside covers/covered. By equals disjoint starts finishes precedes succeeds Fig. 5. 4. Icons for the various synchronization relationships Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 5
Fig. 5. 5. A conceptual schema for a temporal data warehouse Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 6
(a) Temporal level (b) Temporal level with temporal attributes (c) Non-temporal level with temporal attributes Fig. 5. 6. Types of temporal support for a level Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 7
Fig. 5. 7. A nontemporal relationship between temporal levels Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 8
(a) (b) Fig. 5. 8. An example of an incorrect analysis scenario when a nontemporal relationship between temporal levels changes Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 9
Fig. 5. 9. A temporal relationship between nontemporal levels Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 10
(a) (b) Fig. 5. 10. Links kept by a temporal relationship between nontemporal levels: (a) before and (b) after deleting a section Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 11
Fig. 5. 11. A temporal relationship between temporal levels Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 12
Fig. 5. 12. Instant and lifespan cardinalities between hierarchy levels Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 13
Fig. 5. 13. Schema for analysis of indemnities paid by an insurance company Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 14
Fig. 5. 14. Inclusion of loading time for measures Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 15
(a) Events (b) States Fig. 5. 15. Inclusion of valid time for measures Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 16
Fig. 5. 16. Usefulness of including both valid time and loading time Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 17
Fig. 5. 17. A temporal data warehouse schema for an insurance company Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 18
Fig. 5. 18. Usefulness of valid time, transaction time, and loading time Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 19
(a) (b) (c) Fig. 5. 19. An example of distribution of measures in the case of temporal relationships Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 20
Fig. 5. 20. Different temporal granularities in dimensions and measures Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 21
Fig. 5. 21. Example of a coercion function for salary Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 22
Fig. 5. 22. Metamodel of the temporally extended Multi. Dim model Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 23
(a) ER schema (b) Object-relational representation (c) ER schema (d) Object-relational representation Fig. 5. 23. Mapping levels with temporal attributes Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 24
(a) ER schema (b) Object-relational representation Fig. 5. 24. Mapping a temporal level Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 25
(a) ER schema (b) Object-relational representation Fig. 5. 25. Mapping a hierarchy with a nontemporal relationship Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 26
(a) (b) (c) Fig. 5. 26. Various cardinalities for temporal relationships linking nontemporal levels Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 27
Fig. 5. 27. Mapping the schema given in Fig. 5. 26 a into the ER model Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 28
(a) One-to-many cardinalities (b) One-to-many cardinalities (c) Many-to-many cardinalities Fig. 5. 28. Various object-relational representations of temporal links Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 29
(a) Multi. Dim schema (b) Object-relational representation Fig. 5. 29. Mapping a temporal relationship between temporal levels Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 30
(a) ER representation (b) Relational table for the Quantity measure (c) Object-relational representation Fig. 5. 30. Mapping of the fact relationship shown in Fig. 5. 5 Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 31
(a) Excerpt from the schema shown in Fig. 5. 5 Fig. 5. 31. Approaches to measure aggregation in the presence of temporal relationships Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 32
(b) Tables for the parent-child and fact relationships (c) Sales districts and the measure Quantity for stores S 1, S 2, and S 3 (d) Traditional temporal aggregation (e) Eder and Koncilia’s aggregation Fig. 5. 31. Approaches to measure aggregation in the presence of temporal relationships Copyright © 2008 Elzbieta Malinowski & Esteban Zimányi 33
- Slides: 33