SQL in a Nutshell, 2nd Edition

Identifiers

Keep in mind that RDBMSs are built upon set theory. In ANSI terms, clusters contain sets of catalogs, catalogs contain sets of schemas, schemas contain sets of objects, and so on. Most database platforms use corollary terms: instances contain one or more databases; databases contain one or more schemas; schemas contain one or more tables, views, or stored procedures, and the privileges associated with each object. At each level of this structure, items require a unique name (that is, an identifier) so that they can be referenced by programs and system processes.This means that each object (whether a database, table, view, column, index, key, trigger, stored procedure, or constraint) in a RDBMS must be identified. When issuing the command that creates a database object, you must specify an identifier (i.e., a name) for that new object.

There are two important categories of rules that experienced programmers keep in mind when choosing an identifier for a given item:

Naming conventions

Include logical rules of thumb or naming conventions that ultimately create better database structures and data tracking. These are not so much required by SQL as they are the distilled experience of practiced programmers.

Identifier rules

Are those set by the SQL standard and implemented by the platforms. Such rules include characteristics like how long a name may be. These identifier conventions are covered for each vendor later in this chapter.

Naming conventions

Naming conventions establish ...