This thesis shows a way to represent relational database content as a virtual XML document and how to process queries in natural XML access language (XPath) against this data. Design principals were both, efficiency and minimal system requirements to support also relational database systems without any type of build-in XML support. This is a read-only approach so far. The supported XPath features are

• NodeTest: NameTest and NodeType
• 8 axes (ancestor, ancestor-or-self, attribute, child, descendant, descendant-or-self, parent, self, unrolling is limited to mapping definitions depth)
• relative navigation (starting from context node)
• Constants: Numbers and Literals (Strings)
• Operations: in principle all standard XPath 1.0 operations (additive (+, -), multiplicative (*, div, mod), and, or, equality (=, ≠), relational (>, ≥, <, ≤), union, sign)