State Multi-dimensional Data In Time - A Conceptual Model And Management Strategy Research

Since the 20th century 90's,business activities have become more complex and drastically. Enterprises need to use more strategic decision-making methods to improve their competitiveness. Data warehouse is a new paradigm to provide strategic information.It can provide a number of supports of characteristic decision-making,like complex aggregations,analysis,data comparing and so on,which makes decision-makers increasingly rely on the data warehouse to obtain the historical data to support strategic decision-making organizations.The most important use of Data Warehouse is to do the online analytical processing(OLAP) according to multi-dimensional view of data and gives the data comparing based on the related dimensions in field applications.OLAP system can effectively deal with change of measures.It can consider many measures on many dimensions to accomplish analysis and data slices and set up a view of them.However,without the supporting of multi-temporal data model,data warehouse can only record a snapshot of the state,and the dimension changes in the structure can not be shown.It will also lead to incorrect results in dynamic fields.Since William H.Inmon,the primogenitor of data warehouses,systematically expounded the idea of data warehouse and related theory in 1993,many scholars discuss on how to deal with time-varying changes in the data effectively.But so far,there is not a temporal data widely accepted.The paper provides a multi-temporal data model expansion based on the researches of temporal database.The model can provide four type of support,such as the effective time,matters of time,the life-cycle and loading time.Multidimensional data model is an important model of concept.We add the temporal support to the layer,the layer properties,the level of measures based on it to record changes in the dimensions,which describes business logic exactly.The model above is also a model of concept,and it is an important element to implementation of data warehouse because it can reflect the needs of business.Therefore,logical modeling is the first step to establish temporal in data warehouse theory.In this paper,we first make the researches on the development and theories of temporal data warehouse and analyze the related concept and methods of OLAP.Then,we extend the temporal expansion on the component of multi dimension data model separately,especially on dimensions,levels and measures.After that,we pay attention to the refection from temporal data model to an operable data model with focus on the conversions of temporal type temporal dimension layer,the relationship between father and son in the mapping process in ER models.We provide a prototype with UML here.In the aspects of temporal data warehouse management,the paper summarizes the strategy based on views of substantiation and defines the temporal entity of view.According to it,the paper brings forward a temporal data warehouse framework which can maintain the temporal views automatically and allow users to do some queries.At last,the paper discuss on the maintains of temporal view and calculations of process.Through comparative analysis,the use of temporal data warehouse model in describing the concept data model and the management of temporal view with entities can reduce the complexity of data warehouse operations and improve OLAP processing.This paper will also apply the above theory to a business case to verify the validity of the model.