Java Integration of XQuery - an Information Unit-Oriented Approach

An infrastructure for integrating XQuery into Java systems is described. The infrastructure comprises a new API (XQJPLUS, built on the standard API XQJ) and a tool for Java code generation. The basic idea of the approach is to deliver query results not in terms of query result items, but in terms of “information units”, ready-to-use entities assembled from the result items. The assembly process is guided by control information embedded into the query result, so that the query controls exactly what will be delivered, and in which form. Information units can represent information in a great variety of forms, including many map types and custom objects. An information unit associates its contents with a name, type information and meta data. The information units produced by a query are collected into a special container ("info tray") which offers name-based, intuitive access to the units. The query-specific structure of an info tray may be formally defined by a “tray schema” from which an "info shape" can be generated, a Java class representing a specific kind of info tray and offering compiler checked data access. As an approach to data integration, info trays are gathered into an "info space", a collection of possibly very heterogeneous data which can be addressed in a uniform way by string expressions ("info path" expressions).