Graph Database Indexing Layer for Logic-Based Tree Pattern Matching Over Intensional XML Document Databases

摘要

Most XML query evaluation approaches are based on the technique of tree pattern query matching (TPQ) to find similar occurrences of the query's path and conditions. Mainly, two types of constraints are matched to evaluate a given query, including hierarchical structure constraints and value-based constraints. However, TPQ technique falls short when it comes to matching the logic-based constraints and non-hierarchical relationships between nodes and entities in the XML document and database. In this paper, we overcome this shortage by providing an abstract graph database layer that provides a logic-based graph relational model to inspect and resolve the logics of the query to choose most relevant nodes in the XML document. Only the subtrees of the relevant nodes will be traversed in the document and the other subtrees will be skipped. We propose the application of graph database as an indexing layer that defines conceptual linking between database entities, beside logic-based assertions and constraints to evaluate XML queries over this layer to find most related entities and traverse only their related nodes in the XML document. In addition, we propose a mapping criteria and algorithm between XQuery and Cypher, which is a query language for Neo4j graph database.