Indexing technique for object-oriented XML data

Path repository is proposed as a novel indexing scheme for object-oriented XML data, which is a bi-level tree to represent compactly and precisely the object-oriented XML data and composed of a group level and element level. At the group level, the path repository provides path summaries and class hierarchies, to store the element/attribute OIDs which are the specific information owned by the subtree rooted in index class, while the element/attribute OIDs inherited are stored at upper level to enable early pruning of a large search space. At the element level, the path repository preserves detailed child-parent links so as to access quickly the parent and improve greatly the query processing efficiency. The group-based element reference is used instead of global IDs to enable the heterogeneous XML values to be differentiated according to their groups with similar element values clustered and indexed.