Databases - Practical PostgreSQL
Previous Page	Home	Next Page

CREATE FUNCTION

Name

CREATE FUNCTION -- Defines a new function within the database.

Synopsis

CREATE FUNCTION 
name
 ( [ 
argtype
 [, ...] ] )
       RETURNS 
returntype

       AS '
definition
'
       LANGUAGE '
langname
'
       [ WITH ( 
attribute
 [, ...] ) ]
CREATE FUNCTION 
name
 ( [ 
argtype
 [, ...] ] )
       RETURNS 
returntype

       AS '
obj_file
' [, '
link_symbol
' ]
       LANGUAGE '
langname
'
       [ WITH ( 
attribute
 [, ...] ) ]

Parameters

name

The name of the new function being created.

argtype

The data type of the argument, or arguments, to be accepted by the new function. There are three general input types you may use: base types, complex types, or the special opaque type. The opaque type explicitly allows the function to accept arguments of invalid SQL types. The opaque type is generally used by internal functions, or functions written in internal language such as C, or PL/pgSQL, where the return type is not provided as a standard SQL data type.

returntype

The data type of the value or values returned by the new function. This may be set as a base type, complex type, setof type (a normal data type, prefixed by setof), or the opaque type.

Using the setof modifier determines that the function will return multiple rows worth of data (by default, a function returns only one row). For example, a return type defined as setof integer creates a function that can return more than a single row of integer values.

attribute

An optional function attribute. Valid attributes, as of PostgreSQL 7.1.x, are isstrict and iscacheable.

definition

The definition of the function to create. This is entered as a string, bound by quotes, though its contents vary widely between languages. The exact content of this string may be an internal function name, a SQL statement, or procedural code in a language such as PL/pgSQL.

obj_file [, link_symbol ]

The file that contains the dynamically loadable object code, and the function name in the C source. The link_symbol is only required if the source C function has a name that is different from the intended SQL function name.

langname

The name of the language the new function is written in. Possible values for this parameter are C , SQL , internal , or the name of a procedural language created using the CREATE LANGUAGE command (e.g., plpgsql ). See CREATE LANGUAGE" for further details.

Results

CREATE: The message returned when a function is created successfully.

Description

Use the CREATE FUNCTION command to create a new function in the connected database. Ownership of the function is set to the PostgreSQL user that created it.

Function attributes

iscachable: The iscacheable attribute specifies that the function will always return the same result when passed the same argument values (i.e., calculated results are cached). Such a function does not perform a database lookup or use information not directly present in the parameter list. This option is used by the optimizer to determine whether it is safe to pre-evaluate the result of a function call based on past calls, rather than re-executing the function on cached values for previously passed arguments.
isstrict: The isstrict attribute specifies that the function is strict in its handling of NULL values. This means that whenever the function is passed a NULL argument, it will not operate, and will simply return a NULL value.

Function overloading

PostgreSQL allows function overloading . Users of object-oriented programming languages may be familiar with this term. In PostgreSQL, the term means to create multiple functions with the same name, provided each of them has a unique set of argument types.

Overloading is useful for creating what seems to be a single function that can handle a large variety of different input types; to the user, the series of functions you have created become a single, seamless, versatile tool.

Differing from PostgreSQL's ability to overload functions based on argument types, two compiled C functions in one object file are unable to share the same name. To avoid this problem, you can arbitrarily rename the second C function that you wish to overload within PostgreSQL to a unique function name in your C source, compile the object code, and then explicitly define the link_symbol parameter as that arbitrary name when creating the overloaded C function.

Examples

The following example creates a simple SQL function that returns a book title based on the ID number passed to the function:

booktown=# 
CREATE FUNCTION title(integer) RETURNS text

booktown-# 
       AS 'SELECT title from books where id = $1'

booktown-# 
       LANGUAGE 'sql';

CREATE

The title() function can now be used within the booktown database to retrieve rows with ID numbers matching the number passed as an argument:

booktown=# 
SELECT title(41472) AS book_title;

      book_title
----------------------
 Practical PostgreSQL
(1 row)

Databases - Practical PostgreSQL
Previous Page	Home	Next Page