Stream: implementers
Topic: type in differential
Nick George (Sep 24 2018 at 22:20):
Hello, Hoping to get a bit of clarification on when the ElementDefinition type should appear in a differential. The cases that are confusing to me are when we have a CodeableConcept slicing on a profile, e.g., bodyheight: https://www.hl7.org/fhir/bodyheight.profile.json.html.
Take the elements: Observation.code.coding:bodyheightcode of type Coding, and Observation.code.coding:bodyheightcode.code of type Code. In both cases, the type is unchanged in the profile from the base value defined in Observation,
but the differential in BodyHeight for the Code lists type: [ {code: "code"}]
while the differential for coding does not list the type. Why does it appear for Code but not Coding?
Nick George (Sep 24 2018 at 22:22):
(I should clarify - in neither of those cases is the base type actually defined in Observation, since Observation only defines Observation.code. Both Observation.code.coding and Observation.code.coding.code are defined through the CodeableConcept type
Lloyd McKenzie (Sep 24 2018 at 23:56):
Some of the tooling requires a declaration of type if you're going to assert a fixed value or pattern - that may have driven the inclusion.
Nick George (Sep 25 2018 at 00:09):
Yeah I figured it was something like that. Are rules like that documented anywhere? I've read the slicing and profiling page :)
Lloyd McKenzie (Sep 25 2018 at 00:11):
That's not so much a rule as it is a tooling limitation. Not all tools will necessarily have the limitation
Last updated: Apr 12 2022 at 19:14 UTC
