A uniform mechanism to support long fields and nested relations in database management systems

Abstract

Two seemingly distinct features of next-generation database management systems (DBMSs): long fields and nested relations. A long field contains an uninterpreted sequence of bytes of potentially enormous length. A nested relation is a relation that has relation-valued attributes; the nesting of relations can be arbitrarily deep. By elevating the semantics of long fields from a sequence of bytes to a sequence of instances of a (potentially complex) data type, the mechanisms for storing and retrieving nested relations and long fields are equated. An implementation of these ideas in the context of the GENESIS extensible DBMS is described.<>