Chapter 4 - Routines and Routine Invocation

Abstract

This chapter covers Structured Query Language (SQL): 1999's routine invocation algorithm as it applies to each form of SQL-invoked routine, giving both a “bird's-eye” view and the incredibly tedious details. An SQL-invoked routine can be a procedure, a function, or a method or it can be a routine that is invoked from SQL. The routine itself might be written in SQL in which case it is also an SQL routine; alternatively, it might be written in some other programming language, in which case it is an external routine. Among other factors, this chapter gives overview on SQL's path, how it affects selection of user-defined type for some contexts, and how it affects the routine invocation algorithm on the three forms of SQL-invoked routines. The most concise definition of an SQL-invoked routine is “a routine that is invoked from SQL code.” The phrase SQL-invoked routine has two components. The second component, the word routine, identifies the concept of a “subprogram” that is executed on behalf of another program, usually through the use of some particular syntax that identifies that subprogram. Over the decades, programming languages have used a variety of terminology for different sorts of subprograms, but the programming community appears to have settled on consistent nomenclature, and SQL's terms are in agreement with those of other modern languages. The first component, “SQL-invoked,” identifies subprograms that are executed in response to instructions issued from SQL statements or expressions.