Efficient Processing of Multi-Paradaim Operations on Tree-Structured Data

2014 Fiscal Year Final Research Report

• Project/Area Number: 23650048
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Media informatics/Database
• Research Institution: Kyoto University
• Principal Investigator: TAJIMA Keishi 京都大学, 情報学研究科, 教授 (60283876)
• Project Period (FY): 2011-04-28 – 2015-03-31
• Keywords: 木構造データ / ナビゲーション / 走査 / 検索 / パス式 / 記憶方式 / データ構造 / 処理方式

Outline of Final Research Achievements

There are three paradigms of processing of tree data: processing based on navigation from node to node through edges, processing based on scanning of nodes in some order, e.g., pre-order, and processing based on queries, such as those expressed in path expressions. We developed several methods of efficiently executing operations on tree data including multiple paradigms.
For example, we developed a storage scheme for tree data that achieves efficient execution of both navigation-based operations and query-based operations at the same time. We also developed a scheme for efficiently processing a sequence of operations on tree data including both navigation-based operations and scan-based operations. Our scheme repeatedly transforms the data between a data structure for navigation-based operations and a data structure for scan-based operations. To reduce the number of conversions, it changes the order of operations so that the same kind of operations are contiguous as much as possible.

Free Research Field

データ工学