Data Warehouse In The Supermarket Distribution Decision Support Applications

Data Warehouse (DW) is a new technology in the information spaces. Using DW, we can make the best of existing data resources in corporation, convert mass data into information, and then find valuable knowledge and rules from it. Combined with experiences of decision maker, the results can do good for decision making and boost up competitive power of corporations.This paper first describes the generating background and the developing status of DW in brief. Then some essential theories are expatiated in depth such as the conception of DW, data characteristics, data structure, architecture, granularity, metadata, data partition and so on.This paper emphasizes methods, models and process to build DW. To build DW, there are three main steps: designing and creating database, data migrating, querying and analyzing data. When we design database for data warehouse, information packing is introduced. Information packing includes three data models named information package diagram, star scheme and physical data model, corresponding to conceptual model, logic model and physics model in traditional database. Data migration centralizes the research of verifying, extracting, transforming and loading data. Data query and analysis focuses on On Line Analytical Processing (OLAP) and Data Mining (DM).SQL Server 7.0 puts forward a new solution of DW. This paper introduces the framework of Microsoft DW. To build a DW, many tools were needed. Here some tools provided by SQL Server 7.0 were introduced such as OLAP Services, Data Translation Services (DTS), PivotTable Service and English Query.On the basis of researching the theories and analyzing the operation of chain store, this paper brings forward a solution of DW for the Chain Store Distribution Decision Support System (CSD-DSS).Before building DW for CSD-DSS, this-paper first analyzes data inexisting MIS of chain store and operation of distribution in depth. Then this paper builds information package diagram, star scheme and physical data model step by step. The DW for CSD-DSS, named STARBS, is built.This paper builds MID-STARBS as staging area to migrate data. By using DTS, data from OLAT database is loaded into STARBS after verifying, extracting, transforming.This paper builds STARBS-DSS, an OLAP database by using SQL Server 7.0 OLAP Services. An OLAP cube (Sales) is built as an example. Then it analyzes the cube by Multidimensional Expressions (MDX), some examples are presented.This paper also makes some suggestions to ameliorate the model. The model is built on the base of an ideal status. In fact, there are some disturbing data, which may interfere the model. To wipe off such data and amend the model, this paper uses a way of setting "upper limit". Besides it, different data has different significant. This paper uses a way of setting "weight" to reflect this fact.By using Sybase Power Builder 7.0, a three tier Client/Server software, CSD-DSS is developed as a client OLAP tool. It can help to make scientific decisions in the quantity and frequency of chain store distribution. This system is built by using Microsoft SQL Server 7.0 and based on Windows 98. This paper presents the main function modules. Its main functions are to map out plans for distribution and requirement, to maintain the rule of distribution and requirement, to query and analyze and then display the results.Finally, this paper summarized some pivotal aspects in developing data warehouse, arid some shortages in CSD-DSS are mentioned. Some developing directions of data warehouse in the future are predicted too.