The comparison of SQL, QBE, and DFQL as query languages for relational databases

Abstract

Structure Query Language (SQL) and Query By Example (QBE) are the most widely used query languages for Relational Database Management Systems (RDBMS's). However, both of them have problems concerning ease-of-use issues, especially in expressing universal quantification, specifying complex nested queries, and flexibility and consistency in specifying queries with respect to data retrieval. To alleviate these problems, a new query language called 'DataFlow Query Language' (DFQL) was proposed. This thesis investigates the relative strengths and weaknesses of these three languages. We divide queries into four categories: single-value, set-value, statistical result, and set-count value. In each category, a representative set of queries from each language is specified and compared. Some of the queries specified are logical extensions of the other (already defined) queries, which are used to analyze the query languages' flexibility and consistency in formulating logically related queries. We perform a simple experiment of asking NPS CS students to write a small set of queries in all three languages. Based on the analysis, we conclude that DFQL eliminates the problems of SQL and QBE mentioned above. The relative strengths of DFQL comes mainly from its strict adherence to relational algebra and dataflow-based visuality.