Abstract
Data warehousing is an approach to data integration in which integrated information is stored in a data warehouse for direct querying and analysis. To provide fast access, a data warehouse stores materialized views defined over data from its data sources. As a result, a data warehouse needs to be maintained to keep its contents consistent with the contents of its data sources. Incremental maintenance is generally regarded as a more efficient way to maintain materialized views in a data warehouse; however, it is also technically more difficult. Most of the previous work on incremental maintenance of materialized views is confined to the relational database. Unfortunately, existing approaches to incremental maintenance of relational views cannot be directly applied to object-oriented views. Although incremental maintenance of object-oriented views has been investigated in the past few years, the proposed view maintenance algorithms only work in a centralized environment but not in a data-warehousing environment. In this paper, we study the problem of incremental maintenance of object-oriented views in a data- warehousing environment with multiple distributed data sources. In particular, we address two primary issues. First, we identify which kinds of updates to which classes in the data source may cause changes to an object-oriented view. Second, we propose an incremental view maintenance algorithm that is capable of maintaining an object-oriented view defined over multiple distributed data sources. A prototype system has been implemented and the result of our preliminary performance evaluation demonstrates that our incremental view maintenance algorithm is efficient.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call