This document describes using the COBOL Mapping Editor to extract from a CICS DFHCOMMAREA large buffer program where the layout of the COBOL output parameters is different to the layout of the COBOL input parameters, that is, the WM-LCB-INPUT-BUFFER and the WM-LCB-OUTPUT-BUFFER pointers are pointing to different COBOL layouts (different addresses), one for input and one for output.

• Introduction

• Extracting from a CICS DFHCOMMAREA Large Buffer Program

• Mapping Editor User Interface

• Mapping Editor IDL Interface Mapping Functions

---

Introduction

A DFHCOMMAREA Large Buffer Interface has the structure given below in the linkage section. The field subordinated under DFHCOMMAREA prefixed with WM-LCB describe this structure. The field names themselves can be different, but the COBOL data types (usage clauses) must match exactly. The COBOL server has two fixed interface layout structures: one for input, the other for output.

LINKAGE SECTION.
 01 DFHCOMMAREA.
   10 WM-LCB-MARKER                    PIC X(4).
   10 WM-LCB-INPUT-BUFFER              POINTER.
   10 WM-LCB-INPUT-BUFFER-SIZE         PIC S9(8) BINARY.
   10 WM-LCB-OUTPUT-BUFFER             POINTER.
   10 WM-LCB-OUTPUT-BUFFER-SIZE        PIC S9(8) BINARY.
   10 WM-LCB-FLAGS                     PIC X(1).
      88 WM-LCB-FREE-OUTPUT-BUFFER  VALUE 'F'.
   10 WM-LCB-RESERVED                  PIC X(3).
 01 IN-BUFFER.
    02 OPERATION                       PIC X(1).
    02 OPERAND-1                       PIC S9(9) BINARY.
    02 OPERAND-2                       PIC S9(9) BINARY.
 01 OUT-BUFFER.
    02 FUNCTION-RESULT                 PIC S9(9) BINARY.
  . . .
 PROCEDURE DIVISION USING DFHCOMMAREA.
  . . .
    SET ADDRESS OF IN-BUFFER  TO WM-LCB-INPUT-BUFFER.
    SET ADDRESS OF OUT-BUFFER TO WM-LCB-OUTPUT-BUFFER.
* process the IN-BUFFER and provide result in OUT-BUFFER
    EXEC CICS RETURN.

From a programming point of view, the COBOL server behaves as follows:

Variable	Description
WM-LCB-MARKER	Has eye-catcher "XXXX".
WM-LCB-INPUT-BUFFER	Has pointer to a buffer with COBOL server input parameter data. This buffer is described by a COBOL layout structure.
WM-LCB-INPUT-BUFFER-SIZE	Contains size of COBOL server input parameter data.
WM-LCB-OUTPUT-BUFFER	Has pointer to a buffer with length WM-LCB-OUTPUT-BUFFER-SIZE, where the COBOL server writes its output parameter data. This buffer is described by a COBOL layout structure.
WM-LCB-OUTPUT-BUFFER-SIZE	On input, size of WM-LCB-OUTPUT-BUFFER. On return, the size must match the data length returned in the COBOL layout structure of the WM-LCB-OUTPUT-BUFFER. If the called COBOL server returns variable length data, that is, you have mapped Map OCCURS DEPENDING ON or Set Arrays (Fixed <-> Unbounded), and depending on your runtime architecture, consider the following: CICS Socket Listener (EntireX Adapter or RPC Server) Providing a length considering the actual number of occurences instead of the maximum possible (which was provided on input), reduces network traffic and may improve performance. CICS RPC Server Because in this architecture the marshalling is on-host, there will be no impact on network traffic, even if the provided length is set to the maximum possible number of occurences that was provided on input. If the called COBOL server returns fixed-length data, there is no need to change WM-LCB-OUTPUT-BUFFER-SIZE.
WM-LCB-FLAGS	On return, a value of 'F' in this flag indicates that the called COBOL server allocated an output buffer which had to be released by EntireX

WM-LCB-OUTPUT-BUFFER and WM-LCB-FLAGS are normally not changed by the called COBOL server in this scenario.

If there is a need to return the output data in a different storage, this storage must be allocated by EXEC CICS GETMAIN. Return the new storage address in WM-LCB-OUTPUT-BUFFER. Indicate with WM-LCB-FLAGS='F' that the storage is released (EXEC CICS FREEMAIN) by EntireX.

If you have selected an IDL file and opened the COBOL Mapping Editor with an existing COBOL to IDL mapping, continue with Mapping Editor User Interface.

Extracting from a CICS DFHCOMMAREA Large Buffer Program

This section assumes Input Message same as Output Message is not checked. COBOL output and COBOL input parameters are different, that is, the WM-LCB-OUTPUT-BUFFER (as in the large buffer example above) is set to an address that is different to WM-LCB-INPUT-BUFFER.

If you are extracting IDL from a COBOL source or extending the IDL file by extracting an additional COBOL source with interface type CICS with DFHCOMMAREA large buffer, the Extractor Settings dialog appears (see also Step 4: Define the Extraction Settings and Start Extraction).

Make sure the interface type is correct and check box Input Message same as Output Message is cleared.

Press Next to open the COBOL Mapping Editor.

To select the COBOL interface data items of your COBOL server

1. Add the COBOL data items of the input large buffer to Input Message by using the context menu or toolbar available in the COBOL Source View and COBOL Interface. To do this, locate in the PROCEDURE DIVISION the SET ADDRESS OF <x> TO WM-LCB-INPUT-BUFFER statement. The COBOL data item <x> is the input large buffer you are looking for. See Notes.

2. Add the COBOL data items of the output large buffer to Output Message by using the context menu and toolbars available in the COBOL Interface and IDL Interface. To do this, locate in the PROCEDURE DIVISION the SET ADDRESS OF <y> TO WM-LCB-OUTPUT-BUFFER statement. The COBOL data item <y> is the output large buffer you are looking for. See Notes.

3. Continue with COBOL to IDL Mapping.

Notes:

1. Do not select the pointers in the DFHCOMMAREA pointing to the large buffers, in the example above, WM-LCB-INPUT-BUFFER and WM-LCB-OUTPUT-BUFFER.
2. It is very important to select the right COBOL data items describing the COBOL interface correctly. This means the COBOL data items used as parameters must match in number and in sequence of formats (COBOL usage clause).
3. If your COBOL interface contains REDEFINEs, the first REDEFINE path is offered by default. Check manually whether this is the one you want. If not, correct it. You can select any other REDEFINE path.

The user interface of the COBOL Mapping Editor is described below.

Mapping Editor User Interface

This section assumes you have set the extraction settings as described above. The following areas of the COBOL Mapping Editor user interface are described here:

• COBOL Program Selection

• COBOL Source View

• COBOL to IDL Mapping

For COBOL interface types where COBOL input and COBOL output parameters are different, the user interface of the COBOL Mapping Editor looks like this:

	COBOL Program Selection. Currently selected program with interface type  More info
	COBOL Source View. Contains all related sources for the currently selected COBOL program  More info
	COBOL to IDL Mapping. Tree view of your selected COBOL data items and mapping buttons with which you can map these items to your IDL interface  More info

COBOL Program Selection

The COBOL Program Selection displays the current selected COBOL program with its interface type. If you have extracted more than one COBOL program within the associated IDL file, you can switch to another COBOL program with its mapping by selecting the name in the combo box.

COBOL Source View

All COBOL data items contained in the LINKAGE and WORKING-STORAGE SECTION are offered in a text view for selection. The text view contains all related sources (including copybooks) for the currently selected COBOL program. It is used for selecting data items and retrieving information from the original COBOL sources. The light green bar indicates that the data item is already contained in the COBOL Interface; a dark green bar indicates the data item is selectable and can be added to the COBOL Interface. This section can be collapsed. If you open the Editor with Modify Interface it is collapsed by default. The toolbar provides the following actions:

	Add selected COBOL data item to COBOL Interface as Input Message.
	Add selected COBOL data item to COBOL Interface as Output Message.
	Remove selected COBOL data item from COBOL Interface.
	Remove all COBOL data items from COBOL Interface.
	Reset COBOL Interface to initial state.
	Show dialog to modify COBOL Source Characteristics. Not available for interface type COBOL Converter.
	Show dialog to find text in Source.

The same functionality is also available from the context menu.

COBOL to IDL Mapping

This section covers the following topics:

• COBOL Interface
• Mapping Buttons
• IDL Interface

COBOL Interface

The COBOL Interface shows a tree view of your selected COBOL data items describing the interface of the COBOL server. A context menu is available for the COBOL data items, which provides mapping and other functions. On some COBOL data items, decision icons indicate where particular attention is needed, including mapping icons to visualize the COBOL data type and your current mapping.

The COBOL data item names are derived from the COBOL source from which they were extracted. If your COBOL interface contains parameters without a name (for example, the keyword FILLER is used) those COBOL data items are shown as [FILLER]. See FILLER Pseudo-Parameter.

You can modify the COBOL interface using context menu or toolbar; decision and mapping icons provide additional information.

Context Menu

The context menu on COBOL data items provides the following mapping and other functions, depending on the data item type, the COBOL level and the current mapping.

Map to	A suppressed COBOL data item becomes visible in the IDL interface. Used also to select another REDEFINE path.
Suppress	Suppress unneeded COBOL data items.
Set Constant	Set COBOL data items to constant.
Set Array Mapping	Map an array to a fixed sized or unbounded array.
Set Multiple Possible Output (MPO) Structures	Set COBOL data items where the server program decides the output structure used on return. Specify the set of multiple possible output (MPO) structures and the criteria when a structure is used.
Map to Binary	Map a COBOL data item as IDL parameter of type binary (Bn, BV) to exchange binary data (for example images). See Map to Binary and Revert Binary Mapping under Mapping Editor IDL Interface Mapping Functions.
Revert Binary Mapping	Undo the Map to Binary operation and use the standard mapping.
Remove from COBOL Interface	Remove the data item from the COBOL interface. This also removes the mapped IDL parameter from all IDL interfaces for the current COBOL program. See COBOL Program Selection.

See also Mapping Editor IDL Interface Mapping Functions.

Toolbar

The toolbar offers the following actions:

	Create IDL Interface. Creates a new IDL interface based on the current COBOL interface: all IDL parameters are of IDL direction InOut; no IDL parameters are set to constant; for COBOL REDEFINE, the first REDEFINE path is mapped to IDL; FILLERs are suppressed according to your selection, see Step 4: Define the Extraction Settings and Start Extraction.
	Copy current IDL Interface. Creates a duplicate of the current IDL interface: all modifications such as IDL directions, suppress, selection of REDEFINE paths etc. are kept.
	Remove current IDL Interface.
	Rename current IDL Interface.
	Expand the full tree.
	Collapse the full tree.

See also Map to Multiple IDL Interfaces.

Decision Icons

The decision icons in the first column are set on COBOL data items where particular attention is needed:

This icon visualizes a COBOL REDEFINE. It is essential that you map the correct redefine path for your mapping to In, Out or InOut using the context menu. If you map a REDEFINE path, all other sibling REDEFINE paths are automatically set to "Suppress".

Mapping Icons

The following mapping icons on the COBOL data items indicate your current IDL mapping:

	Scalar parameter, mapped to In.
	Scalar parameter, mapped to Out.
	Group parameter, here mapped to In.
	REDEFINE parameter, here mapped to Out.
	Parameter set to Constant.

Mapping Buttons

The following buttons are available:

Map to  ->

A suppressed COBOL data item becomes visible in the IDL interface. Used also to select another REDEFINE path.

Suppress

See Suppress Unneeded COBOL Data Items.

Set Constant...

See Set COBOL Data Items to Constants.

IDL Interface

If you have mapped the COBOL interface to multiple IDL interfaces, select the IDL interface by choosing the tabs. In the IDL Interface tree view, a context menu is also available with the following possibilities:

• Rename the IDL parameter.

• Ungroup an IDL group parameter. This removes the selected mapped IDL group parameter from the IDL interface and moves up the IDL group elements.

• Remove from COBOL Interface. This also removes the mapped IDL parameter from all IDL interfaces for the current COBOL program. See COBOL Program Selection above.

Mapping Editor IDL Interface Mapping Functions

This section covers the following topics:

• Map to

• Map OCCURS DEPENDING ON

• Map to Multiple IDL Interfaces

• Select REDEFINE Paths

• Suppress Unneeded COBOL Data Items

• Set COBOL Data Items to Constants

• Set Arrays (Fixed <-> Unbounded)

• Set Multiple Possible Output (MPO) Structures

• Map to Binary and Revert Binary Mapping

• Ungroup (Flatten) an IDL Group

Map to

With the Map to functions you make a COBOL data item visible as an IDL parameter in the IDL interface, that is, you design the IDL interface by defining input and output parameters.

To map COBOL data items to IDL interface

1. Go step-by-step through all top-level COBOL data items in the COBOL interface and use the Map to function available in the context menu of the COBOL interface and as mapping button to make a COBOL data item visible as an IDL parameter in the IDL interface:

   

2. Do the same for the output message of the COBOL interface.

Notes:

1. If a COBOL group is mapped, all subordinate COBOL data items are also made visible in the IDL interface.
2. With the inverse function Suppress Unneeded COBOL Data Items (see below) available in the context menu of the COBOL interface and as mapping button, a COBOL data item can be removed from the IDL interface.

Map OCCURS DEPENDING ON

You can make the COBOL ODO subject (here COBOL data item TABLE) of a variable-sized COBOL table (see COBOL Tables with Variable Size - DEPENDING ON Clause) visible as an IDL unbounded group (with maximum). The ODO object (here COBOL data item COUNTER-1) is suppressed and therefore not part of the IDL interface. This is because the number of elements of the IDL unbounded group is already implicitly available. See the following example:

01 COUNTER-1 PIC 99.
01 TABLE OCCURS 1 TO 10 DEPENDING ON COUNTER-1
 02 FIELD1 PIC XX.
 02 FIELD2 PIC 99.

To map OCCURS DEPENDING ON

• Add the COBOL subject (here data item TABLE) and ODO object (here data item COUNTER-1) to the input message or to the output message, wherever they belong. It is important both data items are always together per message direction (input or output).

  

Notes:

1. The ODO subject can be mapped to the IDL interface.
2. The ODO object is always suppressed, but is required to be part of the same message direction (Input Message or Output Message) of the COBOL interface.
3. IDL directions are described in the direction-attribute in attribute-list under Software AG IDL Grammar in the IDL Editor documentation.

Map to Multiple IDL Interfaces

Assume the COBOL server program provides multiple functions or operations, in the following example ADD, SUBRACT, MULTIPLY. Some dispatcher front-end code executes the correct function, for example, depending on a function-code or operation-code parameter:

COBOL snippet: The execution of the different functions ADD, SUBTRACT, MULTIPLY is controlled by the COBOL data item OPERATION. The contents of this decide on the function executed:

. . .

     01 OPERATION                      PIC X(1).
     01 OPERAND1                       PIC S9(9) BINARY.
     01 OPERAND2                       PIC S9(9) BINARY.
     01 FUNCTION-RESULT                PIC S9(9) BINARY.
     . . .
     MOVE 0 TO FUNCTION-RESULT.
     EVALUATE OPERATION
         WHEN "+"
            ADD OPERAND1 OPERAND2
            GIVING FUNCTION-RESULT
         WHEN "-"
            SUBTRACT OPERAND2 FROM OPERAND1
            GIVING FUNCTION-RESULT
         WHEN "*"
            MULTIPLY OPERAND1 BY OPERAND2
            GIVING FUNCTION-RESULT
         WHEN . . .

     END-EVALUATE.
. . .

If you have such a situation, a good approach is to expose each COBOL server program function separately as an IDL program. This gives advantages in further processing. See the following examples, depending on your target endpoint:

• Integration Server
  Instead of having a single adapter service for the EntireX Adapter generated with the Integration Server Wrapper, you have separate adapter services, one for each COBOL function.

• Web service
  Instead of having a web service with a single operation generated with the Web Services Wrapper, you get a web service with multiple operations, one operation for each COBOL function.

• DCOM, Java or .NET
  Instead having a class with a single method generated with the respective wrapper (DCOM | Java | .NET) you get a class with multiple methods, one method for each COBOL function.

To map a COBOL interface to multiple IDL interfaces

1. Select the tab with COBOL to IDL Mapping. For each function, define a separate IDL interface with the toolbar functions or :

   

2. Give the IDL interfaces meaningful names with the toolbar function :

   

3. Define the required constant values to the function-code or operation-code parameter, see Set COBOL Data Items to Constants above:

   

For the delivered Example 1: COBOL Server with Multiple Functions:

• First, for step 1 above: Extract and define 3 separate IDL programs ADD, SUBTRACT, MULTIPLY.

• Second, for step 2 above: Rename them to suitabable names, e.g. 'ADD', 'SUBTRACT', 'MULTIPLY'.

• Third, for step 3 above: Define the constants '+', '-' and '*' to the parameter OPERATION respectively.

• Press Finish to create the following IDL together with a server mapping file. See Server Mapping Files for COBOL.

  library 'EXAMPLE' is
  
   program 'ADD' is
    define data parameter
     1 OPERAND1 (I4) In
     1 OPERAND2 (I4) In
     1 FUNCTION-RESULT (I4) Out
    end-define
  
   program 'SUBTRACT' is
    define data parameter
     1 OPERAND1 (I4) In
     1 OPERAND2 (I4) In
     1 FUNCTION-RESULT (I4) Out
    end-define
  
   program 'MULTIPLY' is
    define data parameter
     1 OPERAND1 (I4) In
     1 OPERAND2 (I4) In
     1 FUNCTION-RESULT (I4) Out
    end-define
  

Notes:

1. The following functions are offered to create further mappings from the COBOL interface, resulting in multiple IDL interfaces (IDL programs).

   Icon	Function	Description
   	Create IDL Interface	Creates a new IDL interface based on the current COBOL interface. All IDL parameters are of IDL direction InOut; no IDL parameters are set to constant; for COBOL REDEFINE, the first REDEFINE path is mapped to IDL; FILLERs are suppressed according to your selection, see Step 4: Define the Extraction Settings and Start Extraction.
   	Copy current IDL Interface	Creates a duplicate of current IDL interface. All modifications such as IDL directions, suppress, selection of REDEFINE paths etc. are kept.
   	Rename current IDL Interface	The default name for the IDL interface is based on the COBOL program name plus appended number. With this function you can give the IDL interface a suitable name.
   	Remove current IDL Interface	Deletes the current IDL interface.

2. With the steps 1 thru 3 described here you can emulate the behavior of function Map to Operation of EntireX version 9.6 and earlier.

Select REDEFINE Paths

For COBOL server programs containing COBOL REDEFINEs, the correct REDEFINE path needs to be chosen for the IDL interface.

To select redefine paths

• Use the Map to function available in the context menu of the COBOL interface and as mapping button to make the COBOL REDEFINE path available in the IDL interface.

  

  Begin with the COBOL REDEFINE defined at the highest level first. Work through all inner COBOL REDEFINE data items, going from higher levels to lower levels.

Notes:

1. Only one REDEFINE path of a COBOL REDEFINE can be mapped to the IDL interface. All COBOL REDEFINE siblings are suppressed.
2. If a REDEFINE path is actively mapped to the IDL interface, all COBOL REDEFINE siblings are suppressed.
3. You can suppress all REDEFINE paths of a COBOL REDEFINE. Simply suppress the active REDEFINE path, see Suppress Unneeded COBOL Data Items above.

Suppress Unneeded COBOL Data Items

COBOL data items without any relevant information can be made invisible in the IDL interface. The IDL interface is simplified - it becomes shorter and tidier. This is useful, for example

• for FILLER data items

• if the consuming RPC client or IS service does not need an Out parameter

• if the COBOL data item is an In parameter and a low value can be provided

If you are using an RPC server such as the z/OS (CICS | Batch) or BS2000 RPC server, the amount of data to be transferred to/from the RPC client is also reduced.

To suppress unneeded COBOL data items

• Use the Suppress function available in the context menu of the COBOL interface and as mapping button to make the COBOL data item invisible in the IDL interface:

  

Notes:

1. The COBOL data item is not part of the IDL interface. It is invisible for consuming RPC clients or IS services.
2. The RPC server or EntireX Adapter provides the COBOL data item to your COBOL server with low value, managing the offset to the next COBOL data item.
3. If a COBOL group is suppressed, all subordinate COBOL data items are suppressed as well.
4. With the inverse function Map to (see above) available in the context menu of the COBOL interface and as a mapping button, a COBOL data item can be made visible in the IDL interface again.

Set COBOL Data Items to Constants

COBOL data items that always require fixed constant values on input to the COBOL interface can be made invisible in the IDL interface and initialized with the required constant values. This is useful for keeping the IDL interface short and tidy. Consuming RPC clients or IS services are not bothered with IDL parameters that always contain constants, such as RECORD-TYPES. This function is often used in conjunction with Map to Multiple IDL Interfaces (see above).

To set COBOL data items to constants

1. Use the Set Constant function available in the context menu of the COBOL interface and as mapping button to define a constant value for a COBOL data item:

   

2. You are prompted with a window to enter the constant value:

   

Notes:

1. The COBOL data item is not part of the IDL interface. It is invisible for consuming RPC clients or IS services.
2. The RPC server or EntireX Adapter provides the defined constant in the COBOL data item to your COBOL server.
3. With the function Map to (see above) available in the context menu of the COBOL interface and as a mapping button, a COBOL data item can be made visible in the IDL interface again.

Set Arrays (Fixed <-> Unbounded)

A COBOL server defines in its interface as the last parameter a COBOL Tables with Fixed Size (fixed-size array). In contrast - as the syntax implies - a variable number of elements is transferred in this fixed-size array (input only, output only or both directions are possible). Array elements at the end of the array are unused. Their content is undefined. The current number of elements is transferred directly or implicitly outside the array. There are multiple options to specify how the receiver calculates the number of array elements.

With this mapping you map the fixed-size array of the COBOL interface with the usage described above to an IDL unbounded array in the IDL interface. A consuming RPC client or IS service can use it then as any other IDL unbounded array.

To set arrays from fixed to unbounded or vice versa

1. Select the COBOL table and use the function Set Array Mapping (fixed<->unbounded) available in the context menu. The following window is displayed:

   

2. Select Unbounded Array and the technique for determining the number of elements.

   

The number of array elements is calculated using one of the following options:

• Large Buffer Length (bytes)
  The COBOL server program inspects WM-LCB-INPUT-BUFFER-SIZE (large buffer length for input) for the request and sets WM-LCB-OUTPUT-BUFFER-SIZE (large buffer length for output) for the reply. To determine the number of array elements, the large buffer length is subtracted first to calculate the array length. The result is then divided by the length of one array element. All lengths are in bytes. The following COBOL snippet shows the reply of a large buffer program. It assumes CONTRACT-BUFFER with fix array PACKETI is the large buffer.

         WORKING-STORAGE SECTION.
           77 II                                      PIC S9(4).
         . . .
         LINKAGE SECTION.
           01 DFHCOMMAREA.
            10 WM-LCB-MARKER                          PIC X(4).
            10 WM-LCB-INPUT-BUFFER                    POINTER.
            10 WM-LCB-INPUT-BUFFER-SIZE               PIC S9(8) BINARY.
            10 WM-LCB-OUTPUT-BUFFER                   POINTER.
            10 WM-LCB-OUTPUT-BUFFER-SIZE              PIC S9(8) BINARY.
            10 WM-LCB-FLAGS                           PIC X(1).
               88 WM-LCB-FREE-OUTPUT-BUFFER           VALUE "F".
            10 WM-LCB-RESERVED                        PIC X(3).
           01 CONTRACT-BUFFER.
              04 CONTRACT.
               05 C-ID                                PIC X(8).
               05 C-ACTION                            PIC X(4).
              04 ZONE.
               05 Z-NUMBER                            PIC 9(2).
               05 Z-ID                                PIC X(20).
              04 PACKETI                              OCCURS 99.
               05 P-ID                                PIC X(8).
               05 P-TEXT                              PIC X(30).
               05 P-NUMBER                            PIC 9(2).
          . . .
  
           *  Fill variable output array
              MOVE 0 TO II.
              PERFORM RANDOMNUM TIMES
                 ADD 1 TO II
                 MOVE ... TO P-ID    (II)
                 MOVE ... TO P-TEXT  (II)
                 MOVE ... TO P-NUMBER(II)
              END-PERFORM.
           *  Set large buffer length depending on number of elements
              COMPUTE WM-LCB-OUTPUT-BUFFER-SIZE =
                     (LENGTH OF P-ID  +
                      LENGTH OF P-TEXT   +
                      LENGTH OF P-NUMBER) * II.
              ADD LENGTH OF CONTRACT TO WM-LCB-OUTPUT-BUFFER-SIZE.
              ADD LENGTH OF ZONE     TO WM-LCB-OUTPUT-BUFFER-SIZE.
  
              EXEC CICS RETURN END-EXEC.

• COBOL data item contains array length (bytes)
  The COBOL server program inspects a numeric COBOL data item (ZONED, PACKED or BINARY COBOL type) for the request and sets it accordingly for the reply. This COBOL data item contains the array length. To determine the number of array elements, the contents of the COBOL data item are divided by the length of one array element. All lengths are in bytes. The following COBOL snippet shows how the COBOL interface CONTRACT-DATA is filled by the COBOL server on reply. The length of the fixed-size array PACKETI is contained in COBOL data item C-BYTES.

         WORKING-STORAGE SECTION.
           77 II                                      PIC S9(4).
         . . .
         LINKAGE SECTION.
           01 DFHCOMMAREA.
             03 CONTRACT-DATA.
              04 CONTRACT.
               05 C-ID                                PIC X(8).
               05 C-BYTES                             PIC S9(4).
               05 C-ACTION                            PIC X(4).
              04 ZONE.
               05 Z-NUMBER                            PIC 9(2).
               05 Z-ID                                PIC X(20).
              04 PACKETI                              OCCURS 99.
               05 P-ITEM.
                06 P-ID                               PIC X(8).
                06 P-TEXT                             PIC X(30).
                06 P-NUMBER                           PIC 9(2).
         . . .
  
        *  Fill variable output array
           MOVE 0 TO II.
           PERFORM RANDOMNUM TIMES
              ADD 1 TO II
              MOVE ... TO P-ID    (II)
              MOVE ... TO P-TEXT  (II)
              MOVE ... TO P-NUMBER(II)
           END-PERFORM.
        *  Set table length
           COMPUTE C-BYTES = (LENGTH OF P-ITEM) * II.

• COBOL data item contains length of valid data within messages (bytes)
  The COBOL server program inspects a numeric COBOL data item (ZONED, PACKED or BINARY COBOL type) for the request and sets it accordingly for the reply. To determine the number of array elements, the contents of the COBOL data item are subtracted first to calculate the array length. The result is then divided by the length of one array element. The length of the transferred application data within the message can be shorter than the respective message length. All lengths are in bytes. The following COBOL snippet shows how the COBOL interface CONTRACT is filled by the COBOL server on reply. COBOL data item C-APPDATA contains the length of the valid data of the reply message. The number of array elements of the fixed-size array PACKETI is implicitly contained in COBOL data item C-APPDATA.

          WORKING-STORAGE SECTION.
           77 II                                      PIC S9(4).
           77 EPARM                                   PIC  9(2).
           77 EPARM2                                  PIC  9(4).
         . . .
         LINKAGE SECTION.
           01 DFHCOMMAREA.
              04 CONTRACT.
               05 C-ID                                PIC X(8).
               05 C-APPDATA                           PIC S9(4).
               05 C-ACTION                            PIC X(4).
               05 Z-ID                                PIC X(20).
               05 Z-NUMBER                            PIC 9(2).
              04 PACKETI                              OCCURS 99.
               05 P-ITEM.
                06 P-ID                               PIC X(8).
                06 P-TEXT                             PIC X(30).
                06 P-NUMBER                           PIC 9(2).
         . . .
        *  Fill variable output array
           MOVE 0 TO II.
           PERFORM RANDOMNUM TIMES
              ADD 1 TO II
              MOVE ... TO P-ID    (II)
              MOVE ... TO P-TEXT  (II)
              MOVE ... TO P-NUMBER(II)
           END-PERFORM.
        *  Set length
           COMPUTE C-APPDATA = (LENGTH OF P-ITEM) * II
                             +  LENGTH OF CONTRACT.

• COBOL data item contains number of array elements directly
  The COBOL server program inspects a numeric COBOL data item (ZONED, PACKED or BINARY COBOL type) for the request and sets it accordingly for the reply. The content of the COBOL data item is the number of array elements. The following COBOL snippet shows how the COBOL interface CONTRACT-DATA is filled by the COBOL server on reply. The number of array elements of the fixed-size array PACKETI is directly contained in COBOL data item C-NUM.

         WORKING-STORAGE SECTION.
           77 II                                      PIC S9(4).
         . . .
         LINKAGE SECTION.
           01 DFHCOMMAREA.
             03 CONTRACT-DATA.
              04 CONTRACT.
               05 C-ID                                PIC X(8).
               05 C-NUM                               PIC S9(4).
               05 C-ACTION                            PIC X(4).
              04 ZONE.
               05 Z-NUMBER                            PIC 9(2).
               05 Z-ID                                PIC X(20).
              04 PACKETI                              OCCURS 99.
               05 P-ITEM.
                06 P-ID                               PIC X(8).
                06 P-TEXT                             PIC X(30).
                06 P-NUMBER                           PIC 9(2).
         . . .
  
        *  Fill variable output array
           MOVE 0 TO II.
           PERFORM RANDOMNUM TIMES
              ADD 1 TO II
              MOVE ... TO P-ID    (II)
              MOVE ... TO P-TEXT  (II)
              MOVE ... TO P-NUMBER(II)
           END-PERFORM.
        *  Set occurrences
           MOVE II TO C-NUM.

  Press OK to change the IDL array parameter from fixed array /number to an unbounded array /Vnumber. See array-definition under Software AG IDL Grammar in the IDL Editor documentation. If a COBOL data item is used, it will be set to suppressed because it is superfluous for RPC clients.

  See Suppress Unneeded COBOL Data Items.

Notes:

1. This option should be used carefully and requires knowledge of the COBOL interface. Be aware that an incorrect mapping results in runtime errors.
2. The COBOL table used in this manner must be the last parameter of the COBOL interface; it must not be a subparameter of any other COBOL table and must not contain any DEPENDING ON clause (see COBOL Tables with Variable Size - DEPENDING ON Clause).
3. If a COBOL data item is used, it must be physically located before the COBOL table. The IDL directions must also match.

Set Multiple Possible Output (MPO) Structures

A COBOL server program produces more than one type of output. The layout of the output can therefore take two or more dissimilar shapes. The COBOL server program decides at runtime the output structure returned, that is, the COBOL layout on output varies.

• Multiple Possible Output with REDEFINES
• Optional Output with Groups
• Complex MPO Selections
• MPO Terminology
• Steps

Multiple Possible Output with REDEFINES

A COBOL REDEFINES Clause is often used to describe the possible output structures. In COBOL this is the standard way to describe multiple possible output:

Similar to COBOL data item PAYMENT-DATA in the example below; for this purpose, PAYMENT-DATA is redefined; each redefinition represents an output structure (MPO case); on return exactly one output structure is used; by inspecting COBOL data item PAYMENT-TYPE (MPO selector) first, a caller can determine the returned output structure; the caller then uses the correct redefinition to access the data.

  . . .
    01 INPUT-DATA.
       02 ORDER-NUMBER                            PIC 9(10).

    . . .
    01 OUTPUT-DATA.
        02 <some fields>                          PIC <clause>.
       . . .

        02 PAYMENT-TYPE                           PIC X(2).
           88 PAYMENT-TYPE-VOUCHER                VALUE "VO".
           88 PAYMENT-TYPE-CREDITCARD             VALUE "CC".
           88 PAYMENT-TYPE-TRANSFER               VALUE "TR".
           88 PAYMENT-TYPE-DIRECTDEBIT            VALUE "DB".
          . . .
        02 <preceding data items>                 PIC <clause>.
       . . .
        02 PAYMENT-DATA                           PIC X(256).
        02 PAYMENT-DATA-VOUCHER     REDEFINES PAYMENT-DATA.
           04 VOUCHER-ORIGIN                      PIC X(128).
           04 VOUCHER-SERIES                      PIC X(128).
        02 PAYMENT-DATA-CREDITCARD  REDEFINES PAYMENT-DATA.
           04 CREDITCARD-NUMBER                   PIC 9(18).
           04 CREDITCARD-COMPANY                  PIC X(128).
           04 CREDITCARD-CODE                     PIC 9(12).
           04 CREDITCARD-VALIDITY                 PIC X(8).
        02 PAYMENT-DATA-TRANSFER    REDEFINES PAYMENT-DATA.
           04 TRANSFER-NAME                       PIC X(128).
           04 TRANSFER-IBAN                       PIC X(34).
           04 TRANSFER-BIC                        PIC X(11).
        02 PAYMENT-DATA-DIRECTDEBIT REDEFINES PAYMENT-DATA.
           04 DIRECTDEBIT-IBAN                    PIC X(34).
           04 DIRECTDEBIT-NAME                    PIC X(128).
           04 DIRECTDEBIT-EXPIRES                 PIC 9(8).
          . . .
        02 <subsequent data items>                PIC <clause>.
       . . .

    . . .

*   read order record using ORDER-NUMBER
    . . .

*  set value indicating type of reply (MPO selector)
    IF <some-condition> THEN
      SET PAYMENT-TYPE-VOUCHER TO TRUE
    ELSE IF <some-other-condition> THEN
      SET PAYMENT-TYPE-CREDITCARD TO TRUE
    ELSE IF <some-further-condition> THEN
      SET PAYMENT-TYPE-TRANSFER TO TRUE
    ELSE
      SET PAYMENT-TYPE-DIRECTDEBIT TO TRUE
    END-IF.
   . . .

*   set fields (MPO case) depending on type of reply
    INITIALIZE PAYMENT-DATA.
    EVALUATE TRUE
         WHEN PAYMENT-TYPE-VOUCHER
            MOVE . . . TO VOUCHER-ORIGIN
            MOVE . . . TO VOUCHER-SERIES
         WHEN PAYMENT-TYPE-CREDITCARD
            MOVE . . . TO CREDITCARD-NUMBER
            MOVE . . . TO CREDITCARD-CODE
            MOVE . . . TO CREDITCARD-VALIDITY
         WHEN PAYMENT-TYPE-TRANSFER
            MOVE . . . TO TRANSFER-NAME
            MOVE . . . TO TRANSFER-IBAN
            MOVE . . . TO TRANSFER-BIC
         WHEN PAYMENT-TYPE-DIRECTDEBIT
            MOVE . . . TO DIRECTDEBIT-IBAN
            MOVE . . . TO DIRECTDEBIT-NAME
            MOVE . . . TO DIRECTDEBIT-EXPIRES
         WHEN
            . . .
    END-EVALUATE.
    . . .

In addition, the COBOL interface

• limits the number of possible output structures returned

• defines all possible output structures, that is, they are known during extraction. In the example these are the structures PAYMENT-DATA-VOUCHER, PAYMENT-DATA-CREDITCARD and PAYMENT-DATA-TRANSFER. These are the MPO structures.

• contains an additional COBOL data item carrying a value related to the returned output structure. By inspecting this data item first, the appropriate output structure can be selected to address the data correctly. In the example it is PAYMENT-TYPE. This item is the MPO selector.

• always occupies memory to be able to transfer the longest output structure. If the actual returned output structure is shorter than the longest possible output structure, there is a gap (space) between the multiple possible output and the subsequent data item.

This abstract concept is known as multiple possible output (MPO) EntireX bundles all MPO structures into an MPO group. See MPO Terminology below.

Optional Output with Groups

COBOL group data items can be used to describe optional output structures. The contents of a COBOL data item define under which circumstances COBOL groups are part of the returned data or not. Optional output with group data items are a variant of multiple possible output (MPO).

In addition, the COBOL interface

• limits the number of possible output structures returned

• defines all optional output structures, that is, they are known during extraction. In the COBOL snippet below these are the structures OPTIONAL-OUTPUT-STRUCTURE1 and OPTIONAL-OUTPUT-STRUCTURE2. These are the MPO structures.

• contains an additional COBOL data item carrying an indication which optional output is present. By inspecting this data item first, the appropriate optional output structure can be selected to address the data correctly. If its value does not match, the optional output is not present. In the COBOL snippet it is COBOL data item OPTIONAL-OUTPUT. This item is the MPO selector.

• If the optional output is not present no memory is occupied. There is no gap between the optional output and the subsequent data item, as opposed to Multiple Possible Output with REDEFINES above.

In the COBOL snippet below there are three different shapes of output:

COBOL snippet:

WORKING-STORAGE SECTION.

   01 INPUT-AREA.
      02 FIX-INPUT-ITEM1                      PIC X(4).
      02 <some fields>                        PIC <clause>.
      . . .

   01 OUTPUT-OFFSET                           PIC  S9(9) BINARY.
   01 OUTPUT-AREA                             PIC X(32000).
      . . .

   01 CONTROL-AREA.
      02 OPTIONAL-OUTPUT                      PIC X(1).
         88 OPTIONAL-OUTPUT-1                 VALUE "1".
         88 OPTIONAL-OUTPUT-2                 VALUE "2".
         88 OPTIONAL-OUTPUT-NONE              VALUE "N".
      . . .

   01 OPTIONAL-OUTPUT-STRUCTURE1.
      02 OPTIONAL-OUTPUT-ITEM11               PIC X(10).
      02 OPTIONAL-OUTPUT-ITEM12               PIC X(100).
      02 OPTIONAL-OUTPUT-ITEM13               PIC X(20).
      . . .

   01 OPTIONAL-OUTPUT-STRUCTURE2.
      02 OPTIONAL-OUTPUT-ITEM21               PIC X(4).
      02 OPTIONAL-OUTPUT-ITEM22               PIC X(50).
      02 OPTIONAL-OUTPUT-ITEM23               PIC X(50).
      . . .

   01 FIX-OUTPUT-STRUCTURE1.
      02 FIX-OUTPUT-ITEM11                    PIC X(4).
      02 FIX-OUTPUT-ITEM12                    PIC X(20).
      02 FIX-OUTPUT-ITEM13                    PIC X(8).
      . . .

   01 FIX-OUTPUT-STRUCTURE2.
      02 FIX-OUTPUT-ITEM21                    PIC X(2).
      02 FIX-OUTPUT-ITEM22                    PIC X(10).
      02 FIX-OUTPUT-ITEM23                    PIC X(10).
      . . .


      IF <some-condition> THEN
         SET OPTIONAL-OUTPUT-1 TO TRUE
      ELSE IF <some-other-condition> THEN
         SET OPTIONAL-OUTPUT-2 TO TRUE
      ELSE
         SET OPTIONAL-OUTPUT-NONE TO TRUE
      END-IF.

      . . .

*     provide control area for optional output
      MOVE 1 TO OUTPUT-OFFSET.
      STRING CONTROL-AREA DELIMITED BY SIZE
      INTO OUTPUT-AREA WITH POINTER OUTPUT-OFFSET.

*     provide data items before optional output
      STRING FIX CONTROL-AREA DELIMITED BY SIZE
      INTO OUTPUT-AREA WITH POINTER OUTPUT-OFFSET.

*     provide optional output
      EVALUATE TRUE
          WHEN OPTIONAL-OUTPUT-1
             STRING OPTIONAL-OUTPUT-STRUCTURE1 DELIMITED BY SIZE
             INTO OUTPUT-AREA WITH POINTER OUTPUT-OFFSET
          WHEN OPTIONAL-OUTPUT-2
             STRING OPTIONAL-OUTPUT-STRUCTURE2 DELIMITED BY SIZE
             INTO OUTPUT-AREA WITH POINTER OUTPUT-OFFSET
      END-EVALUATE.

*     provide data items after optional output
      STRING FIX-OUTPUT-STRUCTURE2 DELIMITED BY SIZE
      INTO OUTPUT-AREA WITH POINTER OUTPUT-OFFSET.
      . . .

The returned data is built by copying the necessary COBOL structures into an output area. The optional output is one of OPTIONAL-OUTPUT-STRUCTURE1, OPTIONAL-OUTPUT-STRUCTURE2 or nothing. The presence of the optional output is controlled by a structure named CONTROL-AREA.

Complex MPO Selections

If the MPO case detection is complicated and cannot be defined by available Extractor features (for example the MPO selector and its values), perform the following steps:

To map a complex MPO selection

1. Map the complete MPO group to binary. See Map to Binary and Revert Binary Mapping.

   Note:
   If an MPO group is already defined, you cannot map it to binary. Decide first whether MPO case detection is covered by available extractor features.

2. Implement MPO case detection in your RPC client, using the binary mapping from step 1.

3. Implement MPO case parsing in your RPC client, using the binary mapping from step 1. For the EntireX Adapter, use the COBOL Converter for this purpose. See Converting IS Data Structures with the COBOL Converter in the EntireX Adapter documentation.

MPO Terminology

The following terminology is used with MPOs:

MPO structure

A COBOL group describing the output layout used in an MPO case. All alternative layouts in an MPO group are often described with COBOL REDEFINEs.

MPO group

Bundles together all MPO structures that can be used alternatively. A COBOL interface can contain more than one MPO group.

MPO case

An MPO structure together with its MPO selector values (one or more).

MPO selector

A COBOL data item containing a specific value (MPO selector value) where the actual MPO case can be determined.

• For MPOs based on REDEFINEs, the MPO selector can be placed before, inside or after the MPO group.

• For optional output with groups, the MPO selector precedes the MPO group and is located outside the MPO group.

• Only for MPP Message Interface (IMS Connect): Instead of determining the position of the MPO selector from beginning of the message, you can calculate the position using a fixed offset starting from the end of the message. This alternative is limited to one MPO group per program. See check box MPO Selector determined from message end in step Create a new MPO group below.

MPO selector value

Each value indicates exactly one output structure. An output structure can be indicated by further values.

Steps

To set multiple possible output (MPO) structures with REDEFINES or groups

Use the Set Multiple Possible Output (MPO) Structures function available in the context menu of the COBOL interface to create new or modify existing MPO groups.

1. Set the top-level COBOL data item where the MPO structures are contained to IDL direction Out. Use the Map to Out function for this purpose:

   

2. From the context menu of the COBOL interface of the COBOL REDEFINE, choose Set Multiple Possible Output (MPO) Structures.

   

3. Set Multiple Possible Output (MPO) Structures into MPO Group.

   

4. Create a new MPO group.

   

5. Set MPO selector values for MPO Structures.

   

   Use the functions to delete and to add MPO selector values:

   

   Notes:

   1. To add multiple MPO selector values per MPO structure, use the function multiple times for the same MPO structure (see value 10 and 20 for structure PAYMENT-DATA).
   2. MPO structures without any MPO selector value are unused and suppressed in the IDL interface value (e.g. see structure PAYMENT-DATA-TRANSFER).
   3. Each MPO selector value must uniquely identify an MPO structure. The same value cannot be used more than once for different MPO structures.
   4. If no defined MPO selector value matches at runtime, an empty MPO group is delivered to the RPC client, that is, none of the MPO cases contain any data. No runtime error is produced.

6. Press Finish to create the following IDL together with a server mapping file. See Server Mapping Files for COBOL.

   library 'PAYMENT' is
   
    program 'PAYMENT' is
     define data parameter
      1 INPUT              In
       2 ORDER-NUMBER             (NU10)
      1 OUTPUT             Out
       2 PAYMENT-TYPE             (A2)
       2 PAYMENT-DATA-MPO  Choice
        3 PAYMENT-DATA            (/V1)
         4 PAYMENT-DATA           (AV256)
        3 PAYMENT-DATA-VOUCHER    (/V1)
         4 VOUCHER-ORIGIN         (AV128)
         4 VOUCHER-SERIES         (AV128)
        3 PAYMENT-DATA-CREDITCARD (/V1)
         4 CREDITCARD-NUMBER      (NU18)
         4 CREDITCARD-CODE        (NU12)
         4 CREDITCARD-VALIDITY    (AV8)
     end-define
   

Map to Binary and Revert Binary Mapping

With such a mapping you allow the COBOL server to deal with binary data (for example images). You can also manage Complex MPO Selections.

The menu entry Map to Binary appears only on COBOL data items were it makes sense, for example in Channel Container interface types it is not allowed to map the container reference itself as binary, but inner items can be mapped as binary. Redefine groups will be handled as a block, that means the largest redefine path or redefine base defines the binary length.

When the binary IDL parameter is selected, all corresponding COBOL data items are selected as well.

To undo the binary mapping, select the root COBOL data item (the first of the selection group) and from the context menu choose Revert Binary Mapping.

Ungroup (Flatten) an IDL Group

This function flattens an IDL group. It removes the selected IDL group from the IDL interface and moves up the IDL group elements. By repeating this function, IDL group elements can be moved to IDL level 1, where the correct IDL direction (IN, OUT) can be set.

Example

A COBOL group PAYMENT-DATA contains input as well as output parameters. ORDER-NUMBER and PAYMENT-TYPE are input; PAYMENT-VALIDITY is an output parameter.

01 PAYMENT-DATA.
  02 ORDER-NUMBER                           PIC 9(10).
  02 PAYMENT-TYPE                           PIC X(2).
    88 PAYMENT-TYPE-VOUCHER                 VALUE "VO".
    88 PAYMENT-TYPE-CREDITCARD              VALUE "CC".
    88 PAYMENT-TYPE-TRANSFER                VALUE "TR".
    88 PAYMENT-TYPE-DIRECTDEBIT             VALUE "DB".
  02 PAYMENT-VALIDITY                       PIC X(256).

IDL directions can be set on IDL level 1 only. Therefore, if PAYMENT-DATA is in the IDL, you need to map it to direction INOUT. It is not possible to set it to direction IN, because the direction of PAYMENT-VALIDITY would be wrong. The same applies if you set PAYMENT-DATA to direction OUT; ORDER-NUMBER and PAYMENT-TYPE would then be wrong.

To solve this conflict

1. Remove the group data item PAYMENT-DATA with function Ungroup from the IDL Parameters context menu.

2. Map ORDER-NUMBER and PAYMENT-TYPE to direction IN. See IDL mapping function Map to.

3. Map PAYMENT-VALIDITY to direction OUT. See IDL mapping function Map to.

Restriction: It is not possible to ungroup COBOL data items with the OCCURS clause (IDL arrays).