Stream: argonaut
Topic: duplicated invariant id?
Yunwei Wang (Dec 26 2021 at 15:15):
I found two us-core-1 invariants in US Core v4.1.0. I thought invariant id has to be unique in an IG but apparently IG publisher think differently. Is that a concern?
Vassil Peytchev (Dec 26 2021 at 16:36):
Sounds like a ballot comment is in order
Eric Haas (Dec 27 2021 at 15:18):
There is no rule that I can find. Agree would be nice to make them sequential in IG, but am reluctant to change invariant id's if already widely used and would cause a lot of headaches for implementers.
Brian Postlethwaite (Dec 27 2021 at 21:39):
Having duplicate ids would be more of a problem for implementers. It would cause ambiguity as to which error was being violated if you were doing a translation based in the id.
Lloyd McKenzie (Jan 03 2022 at 04:30):
Only requirement is that the id is unique in the context of a given StructureDefinition. No rule they have to be unique across StructureDefinitions.
Brett Marquard (Jan 03 2022 at 14:53):
Would be nice to be unique per IG....
Brett Marquard (Jan 03 2022 at 14:53):
@Vassil Peytchev + @Yunwei Wang - Should we change Ids already in use?
Vassil Peytchev (Jan 03 2022 at 15:13):
My comment was on the original post that the two invariants were found in the ballot version. if the use of the same invariant id for two different invariants is already published, I have no strong opinions either way. I agree it is nice if all invariants are unique in the IG, but it may be easier to edit individual structure definitions with an independent numbering of invariants.
Yunwei Wang (Jan 04 2022 at 14:29):
I created a ballot comment for this issue. https://jira.hl7.org/browse/FHIR-34646
Last updated: Apr 12 2022 at 19:14 UTC
