Extending the relational algebra to capture complex objects

Abstract

An important direction in database research for non-standard applications (e.g. engineering or design applications) deals with adequate support for complex objects. Without doubt, the provision of network structures and shared subobjects as well as support for dynamic object definition and appropriate manipulation facilities is urgently needed for natural and accurate modeling as well as for efficient processing of the applications' objects. These concepts are the major concern of the molecule-atom data model (MAD model) and its molecule algebra which is introduced in this paper. They make the model stand out compared to the relational model and even to models limited to hierarchical and statically defined complex objects. By means of the molecule algebra a precise and complete specification of one conceivable kind of complex object processing and its inherent semantics is provided. Furthermore, this algebra is used as a sound basis to express the semantics of the high level query language MOL (molecule query language) that is able to deal with complex objects in a descriptive manner.