Naslov
Integrating Semistructured Data into Data Warehousing System

Autori
Vrdoljak, Boris

Vrsta, podvrsta i kategorija rada
Ocjenski radovi, doktorska disertacija

Fakultet
Fakultet elektrotehnike i računarstva

Mjesto
Zagreb

Datum
14.05

Godina
2004

Stranica
149

Mentor
Skočir, Zoran ; Rizzi, Stefano

Ključne riječi
data warehousing system; data warehouse design; multidimensional data model; semi-structured data; XML (eXtensible Markup Language); query languages; prototype tool for data warehouse design

Sažetak
The dissertation proposes a methodology for designing a data warehouse directly from XML data sources. Data warehouse is a database that collects and integrates historical data from heterogeneous sources in order to support decision making process. Given that XML data is semi-structured, a conceptual design of data warehouses from XML sources brings particular challenges. The dissertation proposes an algorithm for the conceptual design, based on identifying and navigating the functional dependencies in XML sources. The algorithm has been specified in detail using the Unified Modeling Language activity diagrams. The result of the algorithm is a conceptual scheme, designed according to a multidimensional data model. The logical and physical schemes for the warehouse can be derived from the conceptual scheme. A prototype tool for data warehouse design from XML sources has been developed to test and verify the proposed algorithm. A functional architecture describing main functions of the tool has been presented in the dissertation. The tool assists a designer in deriving a conceptual scheme for a data warehouse and enables control of all conceptual design phases through a graphical interface.

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika

POVEZANOST RADA