Datu noliktavu konceptuālo modeļu izstrādes metodes un realizācijas mehānismi

Abstract

Promocijas darba merkis ir izstradat efektivas metodes datu noliktavu konceptualo modelu izveidei un implementacijai un definet datu noliktavu konceptualo modelu izveidei situacijas faktoru kopu, kas nosaka piemerotas metodes izveli vai jaunas metodes izstrades nepieciešamibu. Izstradatas tris metodes datu noliktavu konceptualo modelu izstradei - lietotaju vadita, datu vadita un merku vadita metode. Metožu izstrade lietota metožu inženierijas pieeja. Katrai metodei dota formala reprezentacija un metodes komponentu detalizets apraksts. Veikts izstradato metožu situacijas faktoru salidzinajums. Izstradatas metodes sekmigi aprobetas Latvijas Universitates datu noliktavas projektos, atškirigas izstrades situacijas. Izstradatas ari divas datu noliktavu realizacijas metodes - pieejas kontrolei un leni mainigu faktu glabašanai.The goal of the thesis is to develop effective methods for the development and implementation of conceptual models of data warehouses and to define a set of contingency factors that determine the choice between the usage of an existing method and the necessity to develop a new one. Three new methods are presented - user driven, data driven and goal driven. The method engineering approach is used. Methods are formally represented and their components are described in detail. The discovered contingency factors for all three methods are compared. The new methods are successfully applied in data warehouse projects at the University of Latvia in different development situations. Two new methods for data warehouse implementation are presented – the method for the access control and the method for the storage of slowly changing facts.