FHIR Chat · how to handle resources that do not fit their profiles ? · implementers

Stream: implementers

Topic: how to handle resources that do not fit their profiles ?


view this post on Zulip Georg Fette (Jan 23 2019 at 12:44):

Hello, how should the workflow take place for resources that do not fit the specifications of the server they are provideded by ?
The test server http://test.fhir.org provides the profile Condition and has a StructureDefinition for this profile in which the 'category' is defined with a cardinality of 1. Two of the Condition-resources provided by this server, however, have an array for category instead of the single object.
Do I have to check each resource for validity if it's fitting to the expected profile before I process it ?
Greetings

view this post on Zulip Lloyd McKenzie (Jan 23 2019 at 15:41):

Which version of the server are you looking at? r2, r3 or r4? The rules are different for each

view this post on Zulip Georg Fette (Jan 23 2019 at 15:48):

I am currently looking at r3. Where are the rules for the different versions documented ?

view this post on Zulip Lloyd McKenzie (Jan 23 2019 at 15:57):

The STU3 spec is here: http://hl7.org/fhir/STU3/condition.html. (And it supports a cardinality of 0..* for Condition.category)

view this post on Zulip Lloyd McKenzie (Jan 23 2019 at 15:58):

The R4 spec is at http://hl7.org/fhir. The R2 spec is at http://hl7.org/fhir/DSTU2

view this post on Zulip Georg Fette (Jan 23 2019 at 16:10):

yes, I am aware of the general specification of 'Condition' in stu3. With what I have a problem is that the testserver at http://test.fhir.org contains a StructureDefinition resource for the type 'Condition' that overrides the base definition of 0:* with 1:1.

view this post on Zulip Lloyd McKenzie (Jan 23 2019 at 16:21):

Anyone can edit anything on the test server - including changing the StructureDefinitions. However, when I query the definition, it clearly shows 0..*
http://test.fhir.org/r3/StructureDefinition?_id=Condition

view this post on Zulip Lloyd McKenzie (Jan 23 2019 at 16:22):

Are you referring to one of the profiles on Condition? It's totally fine for a profile to tighten the cardinality...

view this post on Zulip Georg Fette (Jan 23 2019 at 16:33):

ah, indeed, I was looking at 'argo-condition'. Sorry for the confusion.
But I am still pondering about the general question how to maintain the conformance between the profiles in the StructureDefinitions and the corresponing resources. Would I have to prevent/forbid changes to the existing StructureDefinition resources ?

view this post on Zulip Lloyd McKenzie (Jan 23 2019 at 16:38):

If you're going to expose profiles on your server that you expect systems to conform to, then yes it makes sense to prohibit them from being updated


Last updated: Apr 12 2022 at 19:14 UTC