Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/19198
標題: 物件導向資料庫系統之靜態雜散架構
Static Hash Schemees for Object-oriented Database Management Systems
作者: 王揮雄
Wang, Hui-Hsung
關鍵字: 物件導向
靜態雜散架構
出版社: 資訊科學研究所
摘要: 在物件導向資料庫系統中,查詢某一個類別之存取範圍可能涉及這個類別之物件,或是它的所有子類別之物件。為有效完成簡單查詢及複雜查詢,其物件存取方法必須能有效支援單一類別存取及階層類別存取。在這篇論文中,我們提出兩種靜態雜散架構:單一類別雜散架構每一類別有一個雜散表;在階層類別雜散架構,所有類別共用一個雜散表,此二架構都只有一個雜散函數。為了達成範圍查詢,選擇嚴格漸增及有界限函數作為雜散函數以保證物件依查詢鍵順序儲存。本論文利用數學方法分析雜散架構,並描述在兩種架構下物件查詢,新建,刪除之演算法。此外,本文亦分析二種架構所需之儲存空間及物件存取效率,並與B+-樹之儲存空間及物件存取效率作比較。
In object-oriented database management systems (OODBMS), the access scope of a query against a class may be the instances of a class or the instances of all its subclasses. To support simple queries and complex queries efficiently, the access method must support both the single class access and the class hierarchy access efficiently. In this thesis, we propose two static hash schemes for OODBMS. The single class hash scheme is that each class has a hash table, while the class hierarchy hash scheme is that only one hash table is used for all classes. There is only one hash function for both hash schemes. To answer the range query, a bounded monotonically increasing function is selected for the hash function to guarantee the search-key order-preserving property. In this thesis, the hash schemes are analyzed mathematicallly, the algorithms for retrieving, inserting, and deleting objects in both schemes are presented, and the access performance and storage cost for both schemes are shown. In addition, the access performance and storage cost for the B+-tree indexes and those for the hash schemes are compared.
URI: http://hdl.handle.net/11455/19198
Appears in Collections:資訊科學與工程學系所

文件中的檔案:

取得全文請前往華藝線上圖書館



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.