For decades, RPG has included the ability to have multiple occurrences of a data structure in a program. Multiple occurrence data structures are defined in RPG IV through the use of the OCCURS keyword on the data structure declaration specification. In RPG III, multiple occurrence data structures are specified by placing the number of occurrences in the "From" columns of the Input specification for the data structure.

In RPG IV, the OCCURS keyword is used to set the number of occurrences, while the OCCUR opcode is used to set or extract the current occurrence. The use of the OCCUR opcode is somewhat clumsy in that you can never address the nth occurrence of a data structure without moving the occurrence index to the desired occurrence and then addressing the data structure subfields.

With OS/400 V5R2, multiple occurrence data structures have been deprecated. The new "Data Structures as Arrays" feature allows you to use the DIM keyword to declare a data structure as an array, effectively giving it multiple occurrences. Figure 1 illustrates the use of the DIM keyword being used to declare a data structure as an array.

Figure 1: A data structure being declared as an array

Subscripting a data structure array is similar to subscripting a regular array index; simply enclose the index in parentheses following the data structure name. The array index is followed by a qualified subfield name. This is the subfield being modified or retrieved. In the example that follows, the subfield CUSTNO of the fifth element (occurrence) of the CONTACT data structure is set to 12345.

Only qualified data structures may be declared as an array. Therefore, the data structure must be declared with one of the QUALFIED, LIKEDS, or LIKEREC keywords. Qualification allows the subfields to be manipulated in do loops or the contemporary for loops. In Figure 2, the for loop iterates through each element of the CONTACT data structure array and sets the value of the BALDUE subfield to zero.

Figure 2: Accessing an array data structure's subfields

The V5R2 Data Structure as Arrays feature makes multiple occurrence data structures much more usable. We no longer need to set the occurrence before accessing the subfields of the occurrence we want. We simply use array subscripting, and we are there.

In Figure 3, I illustrate how multiple occurrence data structures are handled in V5R1 and earlier.

Figure 3: Traditional multiple occurrence data structure

In this example, line 1 declares the CONTACT data structure as a multiple occurrence data structure; the OCCURS(20) keyword is used to declare 20 occurrences. Line 9 in this example is identical to that used in the example in Figure 2. Line 10, however, is where the difference is; it uses the OCCUR operation code to set the current occurrence of the data structure. Finally, on line 11, the BALDUE field is set to zero.

Note that since this is not a qualified data structure, the BALDUE field is not qualified to the parent data structure. If the CONTACT data structure listed in Figure 3 had also included the QUALIFIED keyword, then line 11 would have been specified as follows:

The difference in syntax between this example and what is used in Figure 2 is that the data structure is not an array here but is an array in the example in Figure 2. Therefore, the array scripting parentheses are not specified.

A few things to remember:

• Qualified data structures were introduced in OS/400 V5R1.
• Data structures as arrays were introduced in OS/400 V5R2.
• The LIKEREC keyword (an enhancement to qualified data structures) was introduced in OS/400 V5R2.
• The QUALIFIED, LIKEDS, and LIKEREC keywords cause the data structure to be qualified.
• Only Qualified Data Structures may be declared as arrays.