A template query language for non-first normal form databases

An application of techniques in extended relational databases and models, expert systems, and fuzzy set theory to the problem of data retrieval from a relational database is discussed. A novel approach that permits the retrieval of database information based on a template retrieval language is described. It differs from previous approaches in that retrieval is permitted over a set of tuples that as a group, and not individually, satisfy the query. The retrieval of analogous multituples is complicated by the necessity that the data model and the query language support and use inexact matching techniques. The solution investigated is the use of non-first normal form data models (NFRs). The use of NFRs have several advantages. It is a more natural representation of the underlying data and permits the clustering of similar data. The very nature of the NFR model lends itself to retrieval of analogous multituples. The use of semantic relationships that exist among the attributes is examined, and proper design of NFR relations depending on the above relationships is found to aid in retrieval of analogous multituples