XML Data Warehousing in PostgreSQL

Aaron Sheldon -- Talk was canceled

Contemporary XML parsers store data transiently in memory, and require extensive database development for each XML schema. In the use of large XML based ontologies, and thesauruses the limitations of the DOM, and SAX parsers presents a significant bottleneck in development. But XML itself conforms to a specific syntax, and as such can be generally represented by a fixed table structure in a relational database. The author will demonstrate that through the isomorphism between directed acyclic graphs and the dictionary ordering on integer vectors, a table that records the structure of an XML document in a single column index can be found. Furthermore the single column index allows for queries of ancestry and descendency without recursion. Finally, as well as the database being developed in PostgreSQL, the XML parser was written completely plpgSQL.