Aggregation Strategy And Data Warehouse System Build

Data Warehouse(DW) is higher decision making tools based on analytic data environment. Aggregation incarnates the important character of the integration of DW, offers a type of data query mechanism that is easy to use, fast and consistent response time, provides foundation for OLAP and other applications. So how to construct reasonable and effective aggregation strategy via aggregation metal-data is the emphasis of DW systemThis paper offers the definition of aggregation scheme deepening to dimension level, which provides basis for aggregation strategy, puts forward a type of strategy for aggregation management and navigation based on Tuples. Paper discusses the generating relationship between aggregation schemes, puts forward the way to find optimized path between aggregation schemes. Paper brings forward a user applicable strategy base on DW quality, and advances a way which can delete unnecessary aggregation schemes, so that aggregation strategy can be optimized.DW system is driven by analysis. In this paper, according to inverse life circle of DW, based on the OLAP accessing log, ground on research of aggregation scheme, try to use the autonomy , reactivity and sociality of Agent to refresh and optimize the metadata of DW dynamically, so that administrators can break away from eyeless and onerous job. Author constructs an active aggregation metadata optimization model which is based upon DW and OLAP technology and VisioBroker platform that follows CORBA criterion, which can adjust and optimize aggregation strategy dynamically, build up adapting ability of system.The content of paper is based on scientific research project developed partly by author. SEIDW Studio 1.0 is developed for Bank of China based on MIDAS and middleware technology. SEIDW Studio2.0 is developed for Dalian International Cooperating Group based on open metadata. Paper also gives part of the realization scheme.