花卉批發資訊分享熱線的設計與實作－以台中花市為例

Abstract

台中花卉批發市場的花卉業務情報網裡(Flower Wholesale Information Sharing Hotline, Flower_WISH)已建置完成，而花卉業務情報網架構裡有著複雜的資訊流，這些資訊流的處理可以分成下列三個部份：資料取出/轉換/載入(Extract/Transform/Load, ETL)、資料倉儲、及存取控制。在本篇論文裡，會比較與其它花卉業務情報網的不同點，並詳細討論系統的特色。就ETL而言，是應用Moody-Kortink模型來替代直覺式的資料載入設計，與舊方法比較，這個模型提供了一個較系統化的方法來建構載入的流程。在資料倉儲方面，設計了一個農曆維度加入了看起來像是表格式技術的線上分析系統(On Line Analytical Processing, OLAP)裡，在台灣有許多傳統的慶典節日，故此農曆維度大大的有助於花卉市場來做決策。就存取控制而言，則提出二點特殊的設計：一點針對台中及台南花卉批發市場設計一個共用角色階層，另一點則是設計一個會計/審計的網頁平台系統來監控所有花卉業務情報網的使用者。這二點新增的利益有助於使用者更有效能的來制定決策，亦讓資料流的組織更有效率。

A transaction-based flower Wholesale Information Sharing Hotline (Flower Wholesale Information Sharing Hotline, Flower_WISH) has been designed and implemented for Taichung Flower Wholesale Marketplace. Essentially the complicated concurrent information flows within a flower WISH framework can be divided into three fields: data Extracting / Transforming / Loading (Extract/Transform/Load, ETL), data warehouse, and access control. Compared to other flower WISHs, new salient features are added in this thesis. At the data ETL field, a Moody-Kortink model is applied to replacing the intuitive data loading design. The new approach gives a more systematic way to construct the loading process than the old one. At the data warehouse field, a lunar calendar is designed as an extra dimension in On-Line Analytical Processing (On Line Analytical Processing, OLAP) using look-up table techniques. The lunar calendar dimension is much helpful in making flowers marketing decisions during traditional temple festivals in Taiwan. At the access control field, two distinctive designs are proposed. One is to design a shared role hierarchy for both Taichung and Taiwan Flower Wholesale Marketplaces. The other is to design a web-based accounting/auditing system for monitoring the behavior of WISH users. Two benefits from the new functions added in the WISH are to help the users making decisions more effective and to make the management of information flows more efficient.