Translating between SQL and Tutorial D

Abstract

The relational model of data has a strong influence on the database field and cannot be discarded for any future direction regarding the development of databases (Date and Darwen, 2007). Two relational database languages that are associated with this model are the Structured Query Language (SQL) and Tutorial D. The purpose of this dissertation is to translate between these two languages so that SQL clients can interoperate with a Tutorial D database via SQL. Through studying the literature, it is shown that such translation is important. It will allow the most popular relational database language (i.e. SQL) to interoperate with a true relational database language (i.e. Tutorial D) utilising its features to solve some of the deficiencies that exist in SQL.

The main method conducted for this research is through implementing a translation layer between SQL and Tutorial D using direct string manipulation. The translation logic utilises some of the available methods for string manipulation to manipulate the executed SQL statements and translate them to the corresponding Tutorial D statements. The translation layer then takes care of executing the resultant Tutorial D statements on a chosen Tutorial D database called Rel.

Testing the translation layer consisted of three steps. The first one is to test the translation layer against valid SQL statements to study the feasibility aspect of such translation. It is shown that all the chosen SQL statements for this step pass the test. The second step is to test the translation layer as a solution to the SQL deficiencies defined. It is shown that the test cases that are defined in this step pass the test, and thus, the translation layer solves the deficiencies of SQL. Finally, the translation layer is tested against robustness. All the invalid SQL statements defined in this step are not translated and suitable error messages are shown.

Finally, the dissertation concluded with proving that the translation between SQL and Tutorial D is feasible and can be a suitable approach to solve SQL deficiencies while still using SQL.