Information preservation in XML-to-relational mappings

We study the problem of storing XML documents using relational mappings. We propose a formalization of classes of mapping schemes based on the languages used for defining functions that assign relational databases to XML documents and vice-versa. We also discuss notions of information preservation for mapping schemes; we define lossless mapping schemes as those that preserve the structure and content of the documents, and validating mapping schemes as those in which valid documents can be mapped into legal databases, and all legal databases are (equivalent to) mappings of valid documents. We define one natural class of mapping schemes that captures all mappings in the literature, and show negative results for testing whether such mappings are lossless or validating. Finally, we propose a lossless and validating mapping scheme, and show that it performs well in the presence of updates.