Investigate ways to speed up FHIRJsonParser

FHIRJsonParser is currently generated from FHIR structure definitions using com.ibm.fhir.tools.CodeGenerator. The current structure of the parser is to check for all possible elements which results in more map lookups and method invocations than are needed for a typical instance. A better structure might be something like this:

Patient.Builder builder = Patient.builder();
for (String key : jsonObject.keySet()) {
    switch (key) {
    case "active":
        builder.active(parseBoolean(...));
        break;
    case "gender":
        builder.gender(parseString(...));
        break;
    // ...
}
return builder.build();

[JohnTimm]

Personally I think Deferral of resource deserialization is on the same order of magnitude as #2417 (and I actually like it a bit more). Should we open an issue for that idea?

I have created the following issue: https://github.com/IBM/FHIR/issues/2432

[lmsurpre]

Personally I think Deferral of resource deserialization is on the same order of magnitude as #2417 (and I actually like it a bit more). Should we open an issue for that idea?