SQL$HELP_OLD72.HLB  —  ALTER  FUNCTION
    Allows attributes to be changed for a function that was created
    using the CREATE MODULE statement or the CREATE FUNCTION
    statement.

    It can be used to:

    o  Force a stored (SQL) function to be compiled (COMPILE option)

    o  Modify attributes of external functions

    o  Change the comment on a function

1  –  Environment

    You can use the ALTER FUNCTION statement:

    o  In interactive SQL

    o  Embedded in host language programs

    o  As part of a procedure in an SQL module

    o  In dynamic SQL as a statement to be dynamically executed

2  –  Format

  ALTER FUNCTION <function-name> ++-> COMMENT IS +-> '<string>' -------+++-+->
                                 ||              +------- / <----------+|| |
                                 |+-> COMPILE --------------------------+| |
                                 |+-> NAME <external-body-name> ---------+ |
                                 |+-> external-location-clause ----------+ |
                                 |+-> LANGUAGE language-name ------------+ |
                                 |+-> notify-clause ---------------------+ |
                                 |+-> RETURNS NULL ON NULL INPUT --------+ |
                                 |+-> CALLED ON NULL INPUT --------------+ |
                                 |+-> bind-site-clause ------------------+ |
                                 |+-> bind-scope-clause -----------------+ |
                                 |+-+--------+--+-------------------+----+ |
                                 || +-> NOT -+  +-> VARIANT --------+    | |
                                 ||             +-> DETERMINISTIC --+    | |
                                 |+> RENAME TO <new-function-name> ------+ |
                                 +--------------<--------------------------+

  external-location-clause =

  ---+--> DEFAULT LOCATION ------------+-------------------------+
     +--> LOCATION '<image-location>' -+                         |
   +-------------------------------<-----------------------------+
   +-+----------------------------->--------------------------+-->
     +--> WITH --+-> ALL -----+--> LOGICAL_NAME TRANSLATION --+
                 +-> SYSTEM --+

  notify-clause =

  -> NOTIFY notify-entry-name --> ON -+-+-> BIND ---------+-+->
                                      | +-> CONNECT ------+ |
                                      | +-> TRANSACTION --+ |
                                      +--------- , <--------+

  bind-site-clause =

  --> BIND ON --+--> CLIENT --+--> SITE --->
                +--> SERVER --+

  bind-scope-clause =

  ---> BIND SCOPE --+-> CONNECT ------+-->
                    +-> TRANSACTION --+

3  –  Arguments

3.1  –  bind-site-clause

    Syntax options:

    BIND ON CLIENT SITE | BIND ON SERVER SITE

    Selects the execution model and environment for external routine
    execution.

    CLIENT site binding causes the external routine to be activated
    and executed in the OpenVMS database client (application)
    process. This is the default binding. This binding offers
    the most efficient execution characteristics, allows sharing
    resources such as I/O devices, and allows debugging of external
    routines as if they were part of the client application. However,
    this binding may suffer from address space limitations. Because
    it shares virtual memory with the database buffers, this binding
    is restricted to the client process system user environment, and
    prohibits external routine execution in cases of an application
    running with elevated privileges.

    SERVER site binding causes the external routine to be activated
    in a separate process from the database client and server. The
    process is started on the same node at the database process.
    This binding offers reasonable execution characteristics, a
    larger address space, a true session user environment, and has
    no restrictions regarding client process elevated privileges.
    However, this binding does not permit sharing resources such
    as I/O devices with the client (in particular, there is no
    connection to the client interactive terminal), and debugging
    of routines is generally not possible.

3.2  –  bind-scope-clause

    Syntax options:

    BIND SCOPE CONNECT | BIND SCOPE TRANSACTION

    Defines the scope during which an external routine is activated
    and at what point the external routine is deactivated. The
    default scope is CONNECT.

    o  CONNECT

       An active routine is deactivated when you detach from the
       database (or exit without detaching).

    o  TRANSACTION

       An active routine is deactivated when a transaction is
       terminated (COMMIT or ROLLBACK). In the event that a
       transaction never occurs, the scope reverts to CONNECT.

3.3  –  COMMENT IS string

    Adds a comment about the function. SQL displays the text of the
    comment when it executes a SHOW FUNCTIONS statement. Enclose the
    comment in single quotation marks (') and separate multiple lines
    in a comment with a slash mark (/).

    This clause is equivalent to the COMMENT ON FUNCTION statement.

3.4  –  COMPILE

    The COMPILE option forces the Oracle Rdb server to recompile the
    stored (SQL) function. External functions are not affected.

    Use COMPILE when a function has been made invalid by the
    execution of a DROP . . . CASCADE operation. This mechanism is
    preferred over the SET FLAGS 'VALIDATE_ROUTINE' method available
    in previous versions.

3.5  –  DETERMINISTIC

    Syntax options:

    DETERMINISTIC | NOT DETERMINISTIC

    These clauses are synonyms for the VARIANT and NOT VARIANT
    clauses for conformance to the SQL/PSM standard.

    The DETERMINISTIC clause indicates that the same inputs to the
    function will generate the same output. It is the same as the NOT
    VARIANT clause.

    The NOT DETERMINISTIC clause indicates that the output of the
    function does not depend on the inputs. It is the same as the
    VARIANT clause.

3.6  –  external-body-clause

    Identifies key characteristics of the routine: its name, where
    the executable image of the routine is located, the language in
    which the routine is coded, and so forth.

3.7  –  external-body-name

    The name of the external routine. If you do not specify a name,
    SQL uses the name you specify in the external-routine-name
    clause.

    This name defines the routine entry address that is called for
    each invocation of the routine body. The named routine must exist
    in the external routine image selected by the location clause.

    Unquoted names are converted to uppercase characters.

3.8  –  external-location-clause

    Syntax options:

       DEFAULT LOCATION
       LOCATION 'image-location'

    A default or specific location for the external routine image.
    The resulting file specification must include the type .exe.

    This can be an image file specification or merely a logical name.

    SQL selects a routine based on a combination of factors:

    o  Image string

       The location defaults to DEFAULT LOCATION, which represents
       the file specification string RDB$ROUTINES.

    o  Logical name translation

       The WITH ALL LOGICAL_NAME TRANSLATION and the WITH SYSTEM
       LOGICAL_NAME TRANSLATION clauses specify how logical names in
       the location string are to be translated.

       If no translation option is specified, or if WITH ALL LOGICAL_
       NAME TRANSLATION is specified, logical names are translated in
       the default manner.

       If WITH SYSTEM LOGICAL_NAME TRANSLATION is specified, any
       logical names in the location string are expanded using only
       EXECUTIVE_MODE logical names from the SYSTEM logical name
       table.

3.9  –  LANGUAGE language-name

    The name of the host language in which the external routine
    was coded. You can specify ADA, C, COBOL, FORTRAN, PASCAL, or
    GENERAL. The GENERAL keyword allows you to call routines written
    in any language.

3.10  –  notify-clause

    Specifies the name of a second external routine called (notified)
    when certain external routine or database-related events occur.
    This name defines the routine entry address that is called, for
    each invocation of the notify routine. The named routine must
    exist in the external routine image selected by the location
    clause.

    The events of interest to the notify routine are ON BIND, ON
    CONNECT, and ON TRANSACTION. Multiple events can be specified.

    The following describes the events and scope of each event:

         BIND        Routine activation to routine deactivation
         CONNECT     Database attach to database disconnect
         TRANSACTION Start transaction to commit or roll back
                     transaction

3.11  –  RENAME_TO

    Changes the name of the function being altered. See the RENAME
    for further discussion. If the new name is the name of a synonym
    then an error will be raised.

3.12  –  ON_NULL_INPUT

    Syntax options:

    RETURNS NULL ON NULL INPUT | CALLED ON NULL INPUT

    These clauses control how an external function is invoked when
    one or more of the function arguments is NULL. The CALLED ON
    NULL INPUT clause specifies that the function should be executed
    normally. A normal execution when the PARAMETER STYLE GENERAL
    clause is specified means that SQL should return a run-time error
    when the NULL value is detected.

    The RETURNS NULL ON NULL INPUT clause instructs Oracle Rdb to
    avoid the function call and just return a NULL result. This
    option is valuable for library functions such as SIN, COS,
    CHECKSUM, SOUNDEX, and so on, that usually return an UNKNOWN
    result if an argument is NULL.

    The CALLED ON NULL INPUT clause is the default.

3.13  –  VARIANT

    Syntax options:

    VARIANT | NOT VARIANT

    These clauses are synonyms for the DETERMINISTIC and NOT
    DETERMINISTIC clauses for conformance to the SQL/PSM standard.
    The DETERMINISTIC clause indicates that the same inputs to the
    function will generate the same output. It is the same as the NOT
    VARIANT clause. The NOT DETERMINISTIC clause indicates that the
    output of the function does not depend on the inputs. It is the
    same as the VARIANT clause.

4  –  Examples

    Example 1: Changing a function to be NOT DETERMINISTIC

    When a function is created it is assumed to be DETERMINISTIC.
    That is, given the same input values it should return the same
    result. When a routine has no parameters, such as the GET_TIME
    function shown below, then there is never any variation in the
    input. In this case the function should have been defined as NOT
    DETERMINISTIC to ensure that the Rdb optimizer calls it for each
    row processed, instead of using the previously returned result
    for each row.

    Although DROP FUNCTION and CREATE FUNCTION could have performed
    the same function, ALTER FUNCTION preserves the dependencies that
    exist in the database.

    SQL> alter function GET_TIME
    cont>     not deterministic
    cont>     comment 'Fetch time from clock'
    cont>     /       'Every call must be executed, so change to be'
    cont>     /       'NOT DETERMINISTIC';
    SQL>
    SQL> show function GET_TIME;
    Information for function GET_TIME

     Function is Not Deterministic (variant)
     Function ID is: 262
     External Location is: SYS$SHARE:CLOCKSHR.EXE
     Entry Point is: GET_TIME
     Comment:       Fetch time from clock
                    Every call must be executed, so change to be
                    NOT DETERMINISTIC
     Language is: COBOL
     GENERAL parameter passing style used
     Number of parameters is: 0

    Parameter Name                  Data Type        Domain or Type
    --------------                  ---------        --------------
                                    TIME(2)
            Function result datatype
            Return value is passed by value
Close Help