Modeling and Querying Recursive Data Structures I: Introduction

J.H. ter Bekke and J.A. Bakker (Netherlands)


Query language, recursive data structure, recursive query processing, semantic data modeling, metadata, end user computing.


Present recursive applications for recursive data structures require complex software packages; they cannot be specified in a declarative query language. Recursive queries are complex when a kind of data modeling is applied that emphasizes variable relationships instead of definite and inherent (structural) relationships: users must specify processing details as navigation and iteration. Another kind of modeling supports definite relationships; an example is the relational model. Although this kind of modeling makes control statements in queries superfluous, the relational model still creates problems for end users because relationships between tables cannot be specified inherently; they are specified by relationships between key attributes. Consequently, users interested in data from diverse tables have to specify join operations, which offers opportunities for semantic errors such as joining over non-key attributes. Also semantic data modeling is based on definite relationships, but contrary to the relational model it enables us to specify data structure in an inherent way. As a consequence join terms are superfluous: processing details can be derived by a software system interpreting semantic metadata. Using family trees as an example, we compare the consequences of the two categories of data modeling mentioned above for the specification of non-recursive queries, whereas the following paper [21] will show that inherent specification of data structure is also a fundamental prerequisite for the declarative specification of recursive operations. The resulting processing is reliable and efficient as can be demonstrated by a working DBMS.

Important Links:

Go Back