FHIR Chat · error validation stu3 contained references

Stream: IG creation

Topic: error validation stu3 contained references

Roeland Luykx (Jun 29 2021 at 20:16):

If I make a DocumentManifest with a contained resouce DocumentReference with referenced (also contained - in DocumentManifest) i.e. author and patient, i receive errors with the validator and/or publisher:

Error @ DocumentManifest.contained[0].ofType(DocumentReference).subject (line 16, col25) : ref-1: SHALL have a contained resource if a local reference is provided ( (url: 2; ids: )) [(reference = '#') or reference.startsWith('#').not() or (reference.substring(1).trace('url') in %resource.contained.id.trace('ids'))]
Error @ DocumentManifest.contained[0].ofType(DocumentReference).author[0] (line 20, col18) : ref-1: SHALL have a contained resource if a local reference is provided ( (url: 3; ids: )) [(reference = '#') or reference.startsWith('#').not() or (reference.substring(1).trace('url') in %resource.contained.id.trace('ids'))]

should the actual version of publisher (1.1.76) handle this also correctly or should an older version of publisher be used for stu3? docmanifest.json

Lloyd McKenzie (Jun 29 2021 at 20:38):

STU3 had a bug in one of the invariants that didn't handle certain contained references properly. This was fixed in R4. You may need to live with ignoring the error in STU3. @Grahame Grieve thoughts?

Grahame Grieve (Jun 30 2021 at 02:26):

I don't know. is the invariant in error?

Roeland Luykx (Jun 30 2021 at 04:20):

does it have to do with the content.p[x] 1..1 stuff in stu3 in difference to R4 where content is a Reference?
the difference of the definitions in picture (left R4, right stu3)
image.png

Lloyd McKenzie (Jun 30 2021 at 13:29):

R4 uses %rootResource. STU3 only used %resource.

Roeland Luykx (Jun 30 2021 at 13:38):

Lloyd McKenzie said:

R4 uses %rootResource. STU3 only used %resource.

so it is a definition problem in the STU3 and not in the validator functionality?

Lloyd McKenzie (Jun 30 2021 at 14:39):

Correct

Lloyd McKenzie (Jun 30 2021 at 14:39):

Though we could theoretically put a patch in the validator to work around the STU3 issue. (Not sure what the likelihood of getting a technical correction out on STU3 is at this point.)