RP-Index: Reverse Path Index for XML data

There have been many indexing techniques proposed to support efficient retrieval of XML data in recent years. However, most of these techniques support only a limited number of query types. Some techniques require that there must be a DTD or an XML schema available prior to indexing. In this paper, we propose a novel indexing technique, namely Reverse Path Index (RP-Index), that stores a path to each of XML nodes reversely. RP-Index reduces search space by searching in the collection of unique paths. RP-Index efficiently supports a wide range of query types including branching path queries and regular path queries. RP-Index does not require a prior knowledge about DTD or XML schema. Our experiments show that RP-Index significantly outperforms a traditional approach to indexing of XML in relational databases.