20.2.1. PL/Tcl Functions and Arguments

To create a function in the PL/Tcl language, use the standard syntax

CREATE FUNCTION funcname (argument-types) RETURNS return-type AS '
    # PL/Tcl function body
' LANGUAGE 'pltcl';
     

PL/TclU is the same, except that the language should be specified as pltclu.

The body of the function is simply a piece of Tcl script. When the function is called, the argument values are passed as variables $1 ... $n to the Tcl script. The result is returned from the Tcl code in the usual way, with a return statement. For example, a function returning the greater of two integer values could be defined as:

CREATE FUNCTION tcl_max (integer, integer) RETURNS integer AS '
    if {$1 > $2} {return $1}
    return $2
' LANGUAGE 'pltcl' WITH (isStrict);
     

Note the clause WITH (isStrict), which saves us from having to think about NULL input values: if a NULL is passed, the function will not be called at all, but will just return a NULL result automatically.

In a non-strict function, if the actual value of an argument is NULL, the corresponding $n variable will be set to an empty string. To detect whether a particular argument is NULL, use the function argisnull. For example, suppose that we wanted tcl_max with one null and one non-null argument to return the non-null argument, rather than NULL:

CREATE FUNCTION tcl_max (integer, integer) RETURNS integer AS '
    if {[argisnull 1]} {
        if {[argisnull 2]} { return_null }
        return $2
    }
    if {[argisnull 2]} { return $1 }
    if {$1 > $2} {return $1}
    return $2
' LANGUAGE 'pltcl';
     

As shown above, to return a NULL value from a PL/Tcl function, execute return_null. This can be done whether the function is strict or not.

Composite-type arguments are passed to the procedure as Tcl arrays. The element names of the array are the attribute names of the composite type. If an attribute in the passed row has the NULL value, it will not appear in the array! Here is an example that defines the overpaid_2 function (as found in the older PostgreSQL documentation) in PL/Tcl:

CREATE FUNCTION overpaid_2 (EMP) RETURNS bool AS '
    if {200000.0 < $1(salary)} {
        return "t"
    }
    if {$1(age) < 30 && 100000.0 < $1(salary)} {
        return "t"
    }
    return "f"
' LANGUAGE 'pltcl';
     

There is not currently any support for returning a composite-type result value.

20.2.2. Data Values in PL/Tcl

The argument values supplied to a PL/Tcl function's script are simply the input arguments converted to text form (just as if they had been displayed by a SELECT statement). Conversely, the return command will accept any string that is acceptable input format for the function's declared return type. So, the PL/Tcl programmer can manipulate data values as if they were just text.

20.2.3. Global Data in PL/Tcl

Sometimes it is useful to have some global status data that is held between two calls to a procedure or is shared between different procedures. This is easily done since all PL/Tcl procedures executed in one backend share the same safe Tcl interpreter. So, any global Tcl variable is accessible to all PL/Tcl procedure calls, and will persist for the duration of the SQL client connection. (Note that PL/TclU functions likewise share global data, but they are in a different Tcl interpreter and cannot communicate with PL/Tcl functions.)

To help protect PL/Tcl procedures from unintentionally interfering with each other, a global array is made available to each procedure via the upvar command. The global name of this variable is the procedure's internal name and the local name is GD. It is recommended that GD be used for private status data of a procedure. Use regular Tcl global variables only for values that you specifically intend to be shared among multiple procedures.

An example of using GD appears in the spi_execp example below.

20.2.4. Database Access from PL/Tcl

The following commands are available to access the database from the body of a PL/Tcl procedure:

20.2.5. Trigger Procedures in PL/Tcl

Trigger procedures can be written in PL/Tcl. As is customary in PostgreSQL, a procedure that's to be called as a trigger must be declared as a function with no arguments and a return type of trigger.

The information from the trigger manager is passed to the procedure body in the following variables:

The return value from a trigger procedure can be one of the strings OK or SKIP, or a list as returned by the array get Tcl command. If the return value is OK, the operation (INSERT/UPDATE/DELETE) that fired the trigger will proceed normally. SKIP tells the trigger manager to silently suppress the operation for this row. If a list is returned, it tells PL/Tcl to return a modified row to the trigger manager that will be inserted instead of the one given in $NEW (this works for INSERT/UPDATE only). Needless to say that all this is only meaningful when the trigger is BEFORE and FOR EACH ROW; otherwise the return value is ignored.

Here's a little example trigger procedure that forces an integer value in a table to keep track of the number of updates that are performed on the row. For new rows inserted, the value is initialized to 0 and then incremented on every update operation:

CREATE FUNCTION trigfunc_modcount() RETURNS TRIGGER AS '
    switch $TG_op {
        INSERT {
            set NEW($1) 0
        }
        UPDATE {
            set NEW($1) $OLD($1)
            incr NEW($1)
        }
        default {
            return OK
        }
    }
    return [array get NEW]
' LANGUAGE 'pltcl';

CREATE TABLE mytab (num integer, description text, modcnt integer);

CREATE TRIGGER trig_mytab_modcount BEFORE INSERT OR UPDATE ON mytab
    FOR EACH ROW EXECUTE PROCEDURE trigfunc_modcount('modcnt');
     

Notice that the trigger procedure itself does not know the column name; that's supplied from the trigger arguments. This lets the trigger procedure be re-used with different tables.

20.2.6. Modules and the unknown command

PL/Tcl has support for auto-loading Tcl code when used. It recognizes a special table, pltcl_modules, which is presumed to contain modules of Tcl code. If this table exists, the module unknown is fetched from the table and loaded into the Tcl interpreter immediately after creating the interpreter.

While the unknown module could actually contain any initialization script you need, it normally defines a Tcl "unknown" procedure that is invoked whenever Tcl does not recognize an invoked procedure name. PL/Tcl's standard version of this procedure tries to find a module in pltcl_modules that will define the required procedure. If one is found, it is loaded into the interpreter, and then execution is allowed to proceed with the originally attempted procedure call. A secondary table pltcl_modfuncs provides an index of which functions are defined by which modules, so that the lookup is reasonably quick.

The PostgreSQL distribution includes support scripts to maintain these tables: pltcl_loadmod, pltcl_listmod, pltcl_delmod, as well as source for the standard unknown module share/unknown.pltcl. This module must be loaded into each database initially to support the autoloading mechanism.

The tables pltcl_modules and pltcl_modfuncs must be readable by all, but it is wise to make them owned and writable only by the database administrator.

20.2.7. Tcl Procedure Names

In PostgreSQL, one and the same function name can be used for different functions as long as the number of arguments or their types differ. Tcl, however, requires all procedure names to be distinct. PL/Tcl deals with this by making the internal Tcl procedure names contain the object ID of the procedure's pg_proc row as part of their name. Thus, PostgreSQL functions with the same name and different argument types will be different Tcl procedures too. This is not normally a concern for a PL/Tcl programmer, but it might be visible when debugging.