Enhancement to the PL/I jsonGetValue built-in function to add a selector option to jsonGetValue which would allow specifying a condition for selecting an array element. The function would then only copy those elements that satisfy the condition.

For example, given the JSON document below, it would be great to specify we only want the item with "TRAINING": "Y" copied over to the PL/I structure.

JSON_DOCUMENT:

{
 "PASSES": 3,
 "DATA": [
   { "NAME": "Mather", "ELEVATION": 12100, "TRAINING": "Y" },
   { "NAME": "Pinchot", "ELEVATION": 12130, "TRAINING": "N" },
   { "NAME": "Glenn", "ELEVATION": 11940, "TRAINING": "N" }
 ]
}

PL/I Structure:

dcl
 1 info,
   2 passes fixed bin(31),
   2 data(3),
     3 name char(20) varying,
     3 elevation fixed bin(31)
     3 training char(1);

Usage Example:

read = jsonGetValue(p, n, , selector={data.training = "Y"}, info);

This function call would only return the items with TRAINING="Y" and move them to the PL/I structure.

Such an enhancement would be highly beneficial during parsing. For example, if a web service call returns an array of addresses, and the developer is only interested in the one with “primaryIndicator” set to True. Currently, developers cannot achieve this directly.  In order to do so now, developers must declare a structure and loop through it to find the desired element. The web service does not impose an upper bound on the number of addresses it can return; it merely sends an array and lets the caller pick the relevant one. This is straightforward in other languages like Java, but challenging in PL/I.

We expect this enhancement would be helpful for many use cases.