原生GML数据库存储模型研究

Research on a storage model for native GML databases

针对地理标记语言（GML）数据管理的技术难题和GML对象关系数据库存储冗余度高、空间数据查询效率相对较低的弊端，在分析现有原生GML数据库系统存储特点和方法的基础上，提出了一种面向原生GML数据库的逻辑存储模型——NGDLST模型。基于以元素为单位的细粒度存储组织方式和元素节点区间编码方案，分别定义了组成模型的元素名、属性名、元素值、属性值、元素节点、元素集合和文档集合，设计了模型的数据和索引存储结构。进行了该存储原型系统与三种主流原生XML数据库系统在GML存储与查询方面的对比实验，结果表明该模型是实用和高效的。

In view of the technical difficulties in geography markup language （GML） data management, the high storage redundancy and the low efficient spatial data querying of GML object-relational databases, the paper proposes a log- ical storage model for native GML databases on the basis of the analysis of the existing native GML databases man- agement systems＇ storage characteristics and methods. It is called the NGDLST model. The model＇ s components of element name, attribute name, element value, attribute value, element node, element collection and document col- lection are defined based on a range encoding scheme and a tiny object storage mode which takes an element as the unit to store GML data, and on the basis of this, the model＇ s storing structure for data and indexes is given. The results of the experiment on comparing the performance of the model＇ s prototype system with that of three mainstream native XML database management systems show that the proposed model can perform better in GML storage and query.