| Summary: | 碩士 === 國立交通大學 === 工業工程與管理系所 === 92 === Data warehouse based Wholesale Information Sharing Hotline (WISH) has become more popular in flower industry recently. Also this popularity requests to construct a more efficient data warehouse for satisfying more and more users crowded in on-line. In this thesis, five methods are proposed for improving the performance of data warehouse at Changhua Flower ‘Wholesale Market. First, the measured and dimensional data are transformed separately in Data Transformation Service (DTS) stage for reducing the diagnosis complexity in maintenance. Second, a “week” dimension is added for providing more service on-line. Third, compound and surrogate keys are designed in solving the update problem for input data coming from database without 3rd normal form. Fourth, an exception database is used to store unexpected input data for future processing. Fifth, better hardware and firewall are set up for preventing attacks from hackers. The combination of the proposed methods shows amazing research results not only for improving the performance but also reducing system failures.
|
|---|
