XML Document Indexes: A Classification

Because choosing the most efficient query execution plan relies on indexing techniques, such techniques play an important role in developing query processors. In the Web context, they’re even more crucial, as XML documents are massively used and frequently queried. Given that XML documents are semistructured, however, general query processing techniques — such as those for relational or objectoriented data — won’t work. Researchers have proposed several XMLspecific indexing approaches, but to our knowledge, a unified overview of XML indexing has yet to appear. Such an overview and classification would be useful both theoretically, during design and development, and practically, in helping XML application developers choose the appropriate product. Here, we present a classification of XML indexing techniques based on two key factors. First, we identify the query type that can be optimized. Second, we identify the query processing strategy that benefits from a specific indexing technique. Our classification highlights each technique’s main characteristics, advantages, and drawbacks. In addition to helping developers choose the best XML query management strategies, the classification can help users of those solutions understand why they get specific performance results.