Inserting Keys into the Robust Content-and-Structure (RCAS) Index

Abstract

Semi-structured data is prevalent and typically stored in formats like XML and JSON. The most common type of queries on such data are Content-and-Structure (CAS) queries, and a number of CAS indexes have been developed to speed up these queries. The state-of-the-art is the RCAS index, which properly interleaves content and structure, but does not support insertions of single keys. We propose several insertion techniques that explore the trade-off between insertion and query performance. Our exhaustive experimental evaluation shows that the techniques are efficient and preserve RCAS’s good query performance.