Database and Expert Systems Applications

The talk will discuss several relational models for XMLDatabases and methods, how to map XML-Data to relational data. For each relational model there is a standard technique, how to rewrite XML-Queries in order to transform them into relational SQL-Queries. The relational models and the queries will be considered in combination with classical and multidimensional (UB-tree) indexes and qualitative performance analyses will be presented depending on the relational models and the indexes used. Based on these analyses a recommendation will be given for mapping XML to the relational world. Updatability in Federated Database Systems Mong Li Lee, Sin Yeung Lee, and Tok Wang Ling School of Computing, National University of Singapore email{leeml, jlee, lingtw}@comp.nus.edu.sg Abstract. It is important to support updates in federated database systems. However, not all updates on the federated schema are possible because some may violate certain constraints in the local databases which are involved in the federation. In this paper, we give a formal framework which characterizes the conditions under which a federated schema object type is updatable. We study the steps involved to systematically map an update request on an external view of a federated schema into the equivalent update(s) on the local databases. We also consider the situation where semantically equivalent object types may not model exactly the same set of objects in the real world. We ensure that the set constraints (EQUAL, SUBSET, DISJOINT) between the instance sets of equivalent object types are not violated after an update. It is important to support updates in federated database systems. However, not all updates on the federated schema are possible because some may violate certain constraints in the local databases which are involved in the federation. In this paper, we give a formal framework which characterizes the conditions under which a federated schema object type is updatable. We study the steps involved to systematically map an update request on an external view of a federated schema into the equivalent update(s) on the local databases. We also consider the situation where semantically equivalent object types may not model exactly the same set of objects in the real world. We ensure that the set constraints (EQUAL, SUBSET, DISJOINT) between the instance sets of equivalent object types are not violated after an update.