SQL language summary

Abstract

SQL is a relational database data sublanguage. It is not a complete programming language, but depends on the I/O and control facilities of a host language. It is both a de jure and a de facto standard. ANSI (American National Standards Institute) has published three generations of SQL, as has ISO (International Organization for Standardization). X/Open, a consortium, has also published an SQL specification. Chamberlin and Boyce [1974] published the first paper on what became SQL, based on Codd’s mathematical foundation for logical representation and manipulation of data [Codd 1974]. In 1978, ANSI began to standardize a data definition language for the network database language then being designed by CODASYL; Technical Committee X3H2 was formed for this project, which soon evolved to encompass the entire network database language, published in 1986 as Database Language NDL. X3H2 recognized the importance of the relational model and initiated a project based on Chamberlin’s work. In cooperation with the corresponding ISO group, the SQL specification was developed and published in 1986. SQL-86 omitted support for referential integrity, but SQL-89 added basic referential integrity. SQL-86 and SQL-89 were rightly criticized as inadequate for real applications. In 1992, a major new version, SQL-92, was published, containing features that allowed significant applications without vendor extensions [ANSI 1992; ISO 1992]. SQL has proved key in the success of relational database management systems and is central to many areas, ranging from traditional MIS applications to scientific research. The fourth generation of SQL is currently being prepared. SQL3 adds significant new facilities, including support for object technology, and is partitioned into several parts that can progress independently. See Melton and Simon [1993] for a comprehensive introduction to the SQL language.