An XPath characterization of locally complete XML databases

In this paper, we report ongoing research on what we call locally complete XML databases. These data banks consist of an XML data tree together with a specification of the nodes of the tree that are completely represented in the document. The specification of completeness is defined by means of the defacto standard node selection language for XML, XPath, which makes this approach easy to implement in existing XML/XPath environments. The semantics of locally complete XML databases is given in terms of a set of possible trees. Query answering is defined for certain and possible answers, and its computational complexity is explored.