Building Signature-Trees on Path Signatures in Document Databases

Java is a prevailing implementation platform for XML-based systems. Several high-quality in-memory implementations for the standardized XML-DOM API are available. However, persistency support has not been addressed. In this chapter, we discuss this problem and introduce PDOM (persistent DOM) to accommodate documents as permanent object sets. In addition, we propose a new indexing technique: path signatures to speed up the evaluation of path-oriented queries against document object sets, which is further enhanced by combining the technique of signature-trees with it to expedite scanning of signatures stored in a physical file.