Indexing XML data with a schema graph

I introduce a novel concept of schema graph, which can be used to index XML data. I also describe Xeek, a prototype system based on a relational database, which utilizes my innovation. As in many other methods, the basic idea of Xeek is to decompose the documents into element, attribute, and text nodes, which then are stored into element, attribute, and text relations, respectively. The database schema of Xeek is fixed, so it is possible to store all kinds of XML documents without any prior information about document DTDs. Because Xeek summarizes document structures using a schema graph rather than a schema tree. it is very efficient when querying documents based on their structure. The results of performance studies demonstrate the effectiveness of my method.