Path Summaries and Path Partitioning in Modern XML Databases

The authors study the applicability of XML path summaries in the context of current-day XML databases. They find that summaries provide an excellent basis for optimizing data access methods, which furthermore mixes very well with path-partitioned stores, and with efficient techniques common in today's XML query processors, such as smart node labels (also known as structural identifiers) and structural joins. They provide practical algorithms for building and exploiting summaries, and prove its benefits, alone or in conjunction with a path partitioned store, through extensive experiments.