DEFINE PARAMETER statement
Defines a run-time parameter in an ABL procedure (internal or external), Windows dynamic link library (DLL) routine, UNIX shared library routine, or ActiveX control event procedure.
Note: To define run-time parameters of a user-defined function, or a method within a class (including constructors), see the Parameter definition syntax reference entry.Each parameter requires its own DEFINE statement. The parameters must be specified in the RUN statement in the same order they are defined with DEFINE statements. In addition, the parameter types (INPUT, OUTPUT, INPUT-OUTPUT, RETURN, TABLE, TABLE-HANDLE, DATASET, DATASET-HANDLE, and BUFFER) specified in the DEFINE and RUN statements must agree. The corresponding data types and run-time values must also be compatible enough to allow the AVM to perform any necessary conversions.
Syntax
INPUT PARAMETER
Defines a parameter that gets its value from one of the following sources:
- If the calling procedure runs the current (called) procedure synchronously, the value comes from the corresponding INPUT parameter of the RUN statement.
- If the current procedure is the event procedure specified to handle the PROCEDURE-COMPLETE event for an asynchronous remote procedure, the value comes from the corresponding OUTPUT or INPUT-OUTPUT parameter of the remote procedure.
OUTPUT PARAMETER
Defines a parameter that returns a value to one of the following destinations:
- If the calling procedure runs the current (called) procedure synchronously, the value is returned to the corresponding OUTPUT parameter of the RUN statement in the calling procedure.
- If the calling procedure runs the current (called) procedure as an asynchronous remote procedure, the value is returned to the corresponding INPUT parameter of the event procedure specified to handle the PROCEDURE COMPLETE event for the current procedure.
INPUT-OUTPUT PARAMETER
Defines a parameter that receives an initial value passed from the calling procedure that can be subsequently modified by the called procedure. The calling procedure cannot pass a literal value. The called procedure returns the modified value to one of the following destinations:
- If the calling procedure runs the current (called) procedure synchronously, the value is returned to the corresponding INPUT-OUTPUT parameter of the RUN statement in the calling procedure.
- If the calling procedure runs the current (called) procedure as an asynchronous remote procedure, the value is returned to the corresponding INPUT parameter of the event procedure specified to handle the PROCEDURE COMPLETE event for the current procedure.
RETURN PARAMETERparameter
AS [ HANDLE TO ]primitive-type-name
Specifies a primitive type for the parameter.For ABL procedures,primitive-type-name
can specify any built-in primitive type used to define variables. For more information on the available primitive types, see the Data types reference entry. For more information on defining primitive type variables, see the DEFINE VARIABLE statement reference entry.For DLL or UNIX shared library routines,primitive-type-name
can specify an ABL DLL data type. ABL DLL data types include the built-in ABL data types CHARACTER and MEMPTR, Windows DLL-equivalent data types, and UNIX shared library data types.Table 30 shows how Windows DLL and UNIX shared library data types map to ABL DLL data types.
Table 30: Data types for DLL and UNIX shared library routine parameters Example Cdata type ABL DLL parameterdata type Windows DLL and UNIX shared library data type 8-bit unsigned integer 16-bit signed integer 16-bit unsigned integer long (32-bit UNIX, Win32) int1 LONG2 32-bit signed integer 32-bit unsigned integer__int64
(Win32)long long
(UNIX 32-bit) 64-bit signed integer 4-byte floating point 8-byte floating point Address (usually 32 bits)c-data-type3
HANDLE TOparameter-data-type
3 Address (usually 32 bits)char*
,output-pointer
(which can bechar**
,short**
, and so on), or a pointer to a structure. Address (usually 32 bits)
1The C data typeint
generally specifies a size that depends on the operating system.
2To pass a NULL pointer value to a DLL routine, pass 0 using aLONG
parameter. Do not use a nullMEMPTR
variable to pass a NULL value. However, you can pass a NULL value in one or more elements of aMEMPTR
array. If this conflicts with another way to call the DLL routine, specify a second declaration for the same routine using theORDINAL
option of thePROCEDURE
statement.
3You can use theHANDLE
TO
option to specify a pointer to a primitive type. Therefore, you can use theHANDLE
TO
option with the parameter data types (that is,BYTE
,SHORT
,UNSIGNED-SHORT
,LONG
,FLOAT
, andDOUBLE
) in order to specify a pointer to the respective C data types (that is,char
,short
,unsigned short
,long
,int
,float
, anddouble
). For aCHARACTER
orMEMPTR
parameter, it is redundant because this data type is always passed using a pointer (char*
)
.Caution: ForCHARACTER
parameters, the AVM always passes the routine a pointer to the character or character string value rather than the value itself. If the routine modifies the value, it can also modify the AVM memory outside the bounds of theCHARACTER
value with unpredictable results. For this reason, ABL does not allow you to useOUTPUT
orRETURN
forCHARACTER
orLONGCHAR
parameters, as well asCHARACTER
or LONGCHAR array parameters, and does not recommend you useINPUT-OUTPUT
forCHARACTER
orLONGCHAR
parameters. Rather, pass the character string as aMEMPTR
parameter. For more information, see OpenEdge Development: Programming Interfaces.Note: You cannot useRETURN
for any type of array parameter.To indicate that the DLL or UNIX shared library parameter is a pointer to a value rather than the value itself, use the HANDLE TO option. The HANDLE TO option is required when the DLL routine expects a pointer to the value. Note that the CHARACTER data type implies the HANDLE TO option, whether or not you specify it.For ActiveX control event procedures,primitive-type-name
can specify the built-in ABL data type that maps to the COM object data type of an ActiveX event parameter. Table 31 shows how the COM object data types for event parameters (shown as ActiveX data types) map to ABL data types.
Table 31: Data types for ActiveX control event procedures ActiveX data type1 ABL data type Array ABL array variable Array of bytes RAW Boolean (2-byte integer) LOGICAL Currency (8-byte integer with fixed decimal point) DECIMAL Date DATEDATETIMEDATETIME-TZ Decimal DECIMAL Double (8-byte floating point) DECIMAL Error Code INTEGER Float (Single) DECIMAL Integer (2-byte integer) INTEGER Long (4-byte integer) INTEGER Object (32-bit value) COM-HANDLE String (character string type) CHARACTERLONGCHAR Signed Byte INTEGER Signed Long (4-byte integer) INTEGER Signed Short (2-byte integer) INTEGER Signed 8-byte integer INT64 Unsigned Byte INTEGER Unsigned Long (4-byte integer) INT64 Unsigned Short (2-byte integer) INTEGER Unsigned 4-byte integer INT64 Unsigned 8-byte integer DECIMAL Variant (variable type) <ANYTYPE>2
1For more information on these data type implementations for COM objects, see OpenEdge Development: Programming Interfaces.
2For Variant event parameters, the AppBuilder specifies <ANYTYPE> as a place holder. You must change <ANYTYPE> to the data type that most closely matches the expected value. For more information, see the available documentation on the event parameter.AS [ CLASS ] {object-type-name
}object-type-name
Specifies the type name of an ABL or .NET class or interface. Specify an object type name using the syntax described in the Type-name syntax reference entry. With an appropriate USING statement, you can also specify a class or interface name alone, without the qualifying package or namespace.You cannot directly specify the type name of a .NET mapped object type (such as System.Int32). To define a parameter that matches a .NET mapped type, you must define it as the corresponding ABL primitive type (primitive-type-name
).CLASS
For a class or interface return value, ABL passes an object reference associated with the class or interface, not a class instance itself. For more information on object references, see the Class-based object reference reference entry.LIKE, CASE SENSITIVE, FORMAT, DECIMALS, INITIAL, COLUMN-LABEL, LABEL, NO-UNDO
For descriptions of these options, see the DEFINE VARIABLE statement reference entry.EXTENT [constant
]
Defines the parameter as an array of data elements, where the element data type is specified by the ASprimitive-type-name
option, the LIKEfield
option, or the ASobject-type-name
option. This option can specify an array parameter as either determinate (has a defined number of elements) or indeterminate (has an undefined number of elements). To define a determinate array parameter, specify the EXTENT option with theconstant
argument. This optional argument is an integer value that represents the number of data elements in the array parameter. To define an indeterminate array parameter, specify the EXTENT option without theconstant
argument.The EXTENT is part of the parameter data type. For more information, see the Type-name syntax reference entry.An indeterminate array parameter can be in one of two states: fixed or unfixed, meaning it either has a fixed dimension or it does not. An indeterminate array parameter has an unfixed dimension when first defined. You can fix the dimension of an unfixed indeterminate array parameter by:
- Setting the number of elements in the array parameter using the EXTENT statement
- Defining the indeterminate array parameter so that it becomes the target of a determinate array assignment as a passed argument (on INPUT) or as a value returned from the procedure (on OUTPUT), fixing the indeterminate array to the dimension of the determinate array assignment
ABL treats a fixed indeterminate array parameter as a determinate array parameter; that is, its size is fixed. The AVM determines the size of an unfixed indeterminate array parameter at run time.You cannot pass an unfixed indeterminate array to a COM object, DLL routine, or UNIX shared library routine.If you want to define a parameter that is like an array variable or field, using the LIKE option, but you do not want the parameter to be an array, you can use EXTENT 0 to indicate a non-array parameter.If you are using the AS option and you do not use the EXTENT option (or you specifyconstant
as 0), the parameter is not an array parameter. If you are using the LIKEfield
option and you do not use the EXTENT option, the parameter uses the extent defined for the database field you name (if any).PARAMETER BUFFERbuffer-name
FOR [ TEMP-TABLE ]table-name
[ PRESELECT ]
Defines a buffer parameter, wherebuffer-name
is the name you specify for the buffer andtable-name
is the name of a temp-table or database table. You can pass a buffer associated with a database table to a buffer parameter. You cannot pass a work table to a buffer parameter. A buffer parameter is always INPUT-OUTPUT. You cannot pass buffer parameters to the AppServer.Use the TEMP-TABLE option to define a buffer parameter for a temp-table when the temp-table has the same name as a database table. Otherwise, ABL associates the buffer with the database table by default. Note that you can define a temp-table buffer parameter only for an internal procedure that you define in an external procedure where the temp-table specified bytable-name
is already defined.If you use the PRESELECT option and access the buffer parameter in a DO or REPEAT block, the AVM creates an internal list of the records selected. The PRESELECT option tells the AVM to apply that internal list to the buffer you define.TABLE FORtemp-table-name
Defines a temp-table parameter.You can pass a temp-table parameter to both local and remote procedures. The AVM passes the parameter by value, by default. That is, the caller and the called routine each have their own instance of the temp-table. When you invoke the RUN statement, the AVM deep-copies the parameter from one instance to the other. The table that is copied depends on whether the parameter is INPUT, OUTPUT, or INPUT-OUTPUT. When you pass a temp-table as an INPUT parameter, the AVM replaces the receiving instance with the source instance, by default. You can also append the copied instance to the end of the receiving instance by specifying the APPEND option. For more information about the APPEND option, see the option description later in this reference entry.When passing a temp-table parameter to a local procedure, you can override the default deep copy and pass the parameter by reference or by binding (that is, by specifying the parameter in a RUN statement using either the BY-REFERENCE or BIND option). Passing a temp-table parameter by reference or by binding allows the caller and the called routine to access the same object instance (instead of deep-copying the parameter).Note: When you specify the BIND option in the DEFINE PARAMETER statement, you must also specify the BIND option in the RUN statement.For more information about passing a temp-table parameter by reference or by binding, see the Parameter passing syntax reference entry. For more information about temp-table parameters, see OpenEdge Getting Started: ABL Essentials.TABLE-HANDLEtemp-table-handle
DATASETdataset-name
Defines a compile-time defined ProDataSet object parameter.You can pass a ProDataSet object parameter to both local and remote procedures. The AVM passes the parameter by value, by default. That is, the caller and the called routine each have their own instance of the object. When you invoke the RUN statement, the AVM deep-copies the parameter from one instance to the other. The ProDataSet that is copied depends on whether the parameter is INPUT, OUTPUT, or INPUT-OUTPUT. When you pass a ProDataSet as an INPUT parameter, the AVM replaces the receiving instance with the source instance, by default. You can also append the copied instance to the end of the receiving instance by specifying the APPEND option. For more information about the APPEND option, see the option description later in this reference entry.When passing a ProDataSet object parameter to a local procedure, you can override the default deep copy and pass the parameter by reference or by binding (that is, by specifying the parameter in a RUN statement using either the BY-REFERENCE or BIND option). Passing a ProDataSet object parameter by reference or by binding allows the caller and the called routine to access the same object instance (instead of deep-copying the parameter).Note: When you specify the BIND option in the DEFINE PARAMETER statement, you must also specify the BIND option in the RUN statement.For more information about passing a ProDataSet object parameter by reference or by binding, see the Parameter passing syntax reference entry. For more information on ProDataSet object parameters, see OpenEdge Development: ProDataSets.DATASET-HANDLEdataset-handle
APPEND
Specifies whether or not to append the data from a source instance to the receiving instance of a passed temp-table or ProDataSet parameter. To append input parameter data, specify the APPEND option in the DEFINE PARAMETER statement. To append output parameter data, specify the APPEND option in the RUN statement.BIND
Indicates that a TABLE, TABLE-HANDLE, DATASET, or DATASET-HANDLE parameter binds a reference-only object in one routine to an object instance defined and instantiated in another local routine.When you define a reference-only object in the calling routine, and you want to bind that object definition to an object instance in the called routine, define the parameter by specifying the BIND option in an INPUT or INPUT-OUTPUT parameter definition. When you define a reference-only object in the called routine, and you want to bind that object definition to an object instance in the calling routine, define the parameter by specifying the BIND option in an OUTPUT parameter definition. In either case, the reference-only object definition remains bound to the object instance until the routine containing the reference-only object definition is deleted or terminates.Caution: Do not delete the object or routine to which a reference-only object is bound, or you might be left with references to an object that no longer exists.You can bind multiple reference-only object definitions to the same object instance. You can also bind a single reference-only object definition to the same object instance multiple times without generating an error. However, you cannot bind a single reference-only object definition to multiple object instances.When passing one of these parameters to a remote procedure, the AVM ignores the BIND option and deep-copies the parameter based on the specified parameter mode.For more information about passing these parameters by binding, see the Parameter passing syntax reference entry.BY-VALUE
Specified for an INPUT, OUTPUT, or INPUT-OUTPUT TABLE, TABLE-HANDLE, DATASET, or DATASET-HANDLE parameter in a called routine, this option forces the parameter to be passed to the local routine by value, which overrides any BY-REFERENCE option in the corresponding routine invocation. For more information on BY-REFERENCE, see the Parameter passing syntax reference entry.ExamplesIn the following examples, the
r-runpar.p
procedure runs a subprocedure calledr-param.p
and passes the subprocedure an INPUT parameter. The subprocedurer-param.p
displays the INPUT parameter.
In the following example, the
r-runpr1.p
procedure runs a subprocedure calledr-param1.p
. This example illustrates the use of multiple parameters and shows that the parameters must be passed in the proper order and must be of the same data type. Note that if you do not specify a parameter type in the RUN statement, the AVM assumes it is an input parameter.
In the following example, the
r-runpr2.p
procedure displays information from a database table and assigns the value of a database field to a variable called io-param. The variable is passed as an INPUT-OUTPUT parameter to a subprocedure calledr-param2.p
. The subprocedurer-param2.p
performs a calculation on the INPUT-OUTPUT parameter, then passes it back to the main procedure. Ther-runpr2.p
assigns the value io-param to a database field, then displays io-param.
The following example uses a buffer parameter. The procedure
r-bufp.p
passes the Customer buffer to thegetCustomer
internal procedure, which attempts to find a record using that buffer.
The following example defines parameters for the DLL routine,
MessageBox
, which displays a message on the screen:
Notes
- All procedure parameters are normally passed by value, by default. This means that for any INPUT-OUTPUT or OUTPUT parameter, the field or variable that receives the output value is not set by the called procedure until the procedure returns without error. An exception is made for local DATASET, DATASET-HANDLE, TABLE, and TABLE-HANDLE parameters, which you may pass by reference or by binding by specifying the parameter in a RUN statement using either the BY-REFERENCE or BIND option. If you specify the BIND option in the RUN statement, you must also specify the BIND option in the DEFINE PARAMETER statement.
For more information about passing parameters by reference or by binding, see the Parameter passing syntax reference entry.- You cannot pass a BLOB or CLOB field as a parameter. To pass a BLOB or CLOB field as a parameter, you must include the field in a temp-table or convert the field to its MEMPTR or LONGCHAR counterpart, respectively.
- Buffer parameters are scoped in the same way as shared buffers. They also affect cursors defined in the calling procedure in the same way as shared buffers.
- For DLL or UNIX shared library routine declarations:
- You cannot pass a DATE, DATETIME, or DATETIME-TZ as a parameter to or from a DLL routine or a UNIX shared library routine.
- You can pass a LONGCHAR as a parameter to a DLL routine or a UNIX shared library routine. When passing a LONGCHAR parameter, the AVM passes only the text string (not the code page information). You are responsible for setting the code page of a LONGCHAR parameter.
- You can pass an array of type INTEGER, INT64, or DECIMAL as a parameter to or from a DLL routine or a UNIX shared library routine. You can pass an array of type CHARACTER to (not from) a DLL routine or a UNIX shared library routine.
- You cannot pass a variable or array that contains the Unknown value (
?
) to a DLL.- RETURN parameters are supported only for DLL or UNIX shared library routines. The RETURN parameter type must match the OUTPUT parameter that returns the DLL function value in the RUN statement for the routine. You cannot pass an array as a RETURN parameter to DLL or UNIX shared library routines. Use a MEMPTR instead.
- If you specify a RETURN parameter as MEMPTR to return a character string, use the GET-STRING function to extract the CHARACTER value.
- For more information on DLL routine parameters and how they map to ABL data types, see the chapter on DLLs in OpenEdge Development: Programming Interfaces.
- You cannot pass a MEMPTR as a parameter to or from a COM object.
- You can pass a LONGCHAR, a DATETIME, a DATETIME-TZ, and an array as a parameter to or from a COM object.
When passing a LONGCHAR parameter to a COM object, the AVM passes only the text string (not the code page information). When receiving a text string from a COM object into a LONGCHAR parameter, the AVM converts the text string to the code page associated with the LONGCHAR parameter only if the LONGCHAR has a fixed code page. Otherwise, the AVM sets the LONGCHAR code page to UTF-16. If the AVM cannot convert a LONGCHAR, it raises a run-time error.When passing a DATETIME or DATETIME-TZ parameter to a COM object, the AVM represents the time to the millisecond. When passing a DATETIME-TZ parameter to a COM object, the AVM first converts the DATETIME-TZ value relative to the local session's date and time, and then drops the time zone.When receiving a date from a COM object into a DATETIME or DATETIME-TZ parameter, the AVM represents the time to the millisecond. When receiving a date from a COM object into a DATETIME-TZ parameter, the AVM sets the time zone to the local session's time zone.- For more information on ActiveX event parameters, or using COM objects, see the chapter in OpenEdge Development: Programming Interfaces.
- For dynamic temp-table parameters:
- If the parameter is INPUT TABLE-HANDLE, the temp-table definition behind the handle plus the temp-table contents are sent from the caller to the called routine. The called routine may have either the dynamic INPUT TABLE-HANDLE or the compile-time defined INPUT TABLE as a matching parameter.
- When running remotely with the AppServer, if the parameter is INPUT TABLE-HANDLE, a new instance of the temp-table is created along with its handle, completely separate from the caller’s table, and is populated with the contents from the caller’s table.
You can override this default behavior to allow the calling procedure and the called procedure to access the same object instance by passing the TABLE-HANDLE parameter by reference or by binding (that is, by specifying the parameter in a RUN statement using either the BY-REFERENCE or BIND option). If you specify the BIND option in the RUN statement, you must also specify the BIND option in the DEFINE PARAMETER statement.- The OUTPUT TABLE-HANDLE parameter, just like any other OUTPUT variable, cannot change the data in the caller procedure: in this case, the definition of the temp-table).
- If the parameter is OUTPUT TABLE-HANDLE, the called routine sends back the definition behind the handle along with the contents of the output temp-table. In the caller, if the original handle was the Unknown value (
?
), a new instance of the temp-table is created and populated with the output contents. The temp-table object is created in the default widget-pool of the caller’s OpenEdge session. To avoid memory leaks, the temp-table object must be manually destroyed. For more information on deleting objects, see the reference entry for the DELETE OBJECT statement.If the original handle is not the Unknown value (?
), the caller’s existing table must match the table being received from the called routine.- If the APPEND option is used, the new data is added to the existing table’s data.
- If the parameter is INPUT-OUTPUT TABLE-HANDLE, a combination of the above occurs.
- If you call a remote procedure asynchronously (using the ASYNCHRONOUS option of the RUN statement) and pass a parameter as OUTPUT TABLE-HANDLE
temp-table-handle
APPEND, the event procedure must specify a corresponding DEFINE INPUT PARAMETER TABLE-HANDLE FORtemp-table-handle
APPEND statement, andtemp-table-handle
must be global to both the calling procedure and the event procedure.- If you define an INPUT TABLE parameter for an asynchronous event procedure with a data type that is different from the data type of the corresponding OUTPUT TABLE parameter passed from the AppServer, any failure to convert the passed value causes the event procedure to fail and the AVM to display an error message on the client.
- The AVM creates the temp-table object for an OUTPUT TABLE-HANDLE parameter in the session’s default widget pool. Remember to delete this object at the appropriate point to avoid memory leaks.
- You cannot specify a ProDataSet object or ProDataSet object handle as a parameter for an asynchronous remote procedure.
- For more information on working with asynchronous remote procedures and event procedures, see OpenEdge Application Server: Developing AppServer Applications.
- For dynamic ProDataSet object parameters:
- If the parameter is INPUT DATASET-HANDLE, the ProDataSet object definition behind the handle plus the ProDataSet object contents are sent from the caller to the called routine. The called routine may have either the dynamic INPUT DATASET-HANDLE or the compile-time defined INPUT DATASET as a matching parameter.
- If the parameter is INPUT DATASET-HANDLE, a new instance of the ProDataSet object is created along with its handle, completely separate from the caller’s table, and is populated with the contents from the caller’s table.
You can override this default behavior to allow the calling procedure and the called procedure to access the same object instance by passing the DATASET-HANDLE parameter by reference or by binding (that is, by specifying the parameter in a RUN statement using either the BY-REFERENCE or BIND option). If you specify the BIND option in the RUN statement, you must also specify the BIND option in the DEFINE PARAMETER statement.- If the parameter is OUTPUT DATASET-HANDLE, the handle plus the definition behind the handle are sent from the caller to the called routine. The called routine may have either the dynamic OUTPUT DATASET-HANDLE or the compile-time defined OUTPUT DATASET as a matching parameter.
If the parameter is OUTPUT DATASET-HANDLE, and the caller’s handle is the Unknown value (?
), no definition is sent to the called routine.- If the parameter is OUTPUT DATASET-HANDLE, the called routine sends back the definition behind the handle along with the contents of the output ProDataSet object. In the caller, if the original handle is the Unknown value (
?
), a new instance of the ProDataSet object is created and populated with the output contents. If the original handle is not the Unknown value (?
), the caller’s existing object must match the object being received from the called routine.- If the APPEND option is used, the new data is added to the existing object’s data.
- If the parameter is INPUT-OUTPUT DATASET-HANDLE, a combination of the above occurs.
- A BUFFER parameter cannot be a REFERENCE-ONLY buffer or a BEFORE-TABLE buffer. The use of them as a BUFFER type of parameter (as opposed to INPUT or OUTPUT) is a compiler error. Instead of passing the buffer, the buffer's table can be passed by-reference.
See alsoDEFINE BUFFER statement, DEFINE VARIABLE statement, DELETE PROCEDURE statement, Parameter passing syntax, RUN statement, Type-name syntax, USING statement
OpenEdge Release 10.2B
|