Stream: conformance
Topic: Validating Quantity with code binding
Patrick Werner (Dec 04 2019 at 13:43):
(deleted)
Patrick Werner (May 05 2020 at 15:05):
Just tried to validate an example against a Observation profile with a constrained valueQuantity.
Patrick Werner (May 05 2020 at 15:31):
Error @ Observation.value.ofType(Quantity).code (line 24, col21) : The value provided ("cm") is not in the value set http://fhir.de/ValueSet/einheiten-koerpergroesse (http://fhir.de/ValueSet/einheiten-koerpergroesse), and a code is required from this value set) (error message = Unknown Code http://unitsofmeasure.org#cm in http://unitsofmeasure.org)
Lloyd McKenzie (May 05 2020 at 15:37):
@Grahame Grieve
Patrick Werner (May 05 2020 at 15:39):
currently debugging..
Patrick Werner (May 05 2020 at 15:39):
will provide the example and profile
Patrick Werner (May 05 2020 at 15:59):
problem is that cs = findSpecialCodeSystem(system);returns the wrong CS
Patrick Werner (May 05 2020 at 16:00):
(deleted)
Patrick Werner (May 05 2020 at 16:23):
the CanonicalResourceManager has a wrong key in its map.
Patrick Werner (May 05 2020 at 16:23):
for http://unitsofmeasure.orga CS is returned wich has the url: http://fhir.de/CodeSystem/dimdi/ops
- url/map.key is
http://unitsofmeasure.orgbut the content is from the OPS CS
Patrick Werner (May 05 2020 at 16:23):
(deleted)
Grahame Grieve (May 05 2020 at 21:12):
sounds like a bad package?
Patrick Werner (May 06 2020 at 06:17):
Yes that was the reason
Patrick Werner (May 06 2020 at 06:18):
The validator tried to validate against a CS with content: not-present.
Patrick Werner (May 06 2020 at 06:20):
And then couldn‘t find the code. Would have expected a warning that the code couldn‘t he validated against an empty CS
Grahame Grieve (May 06 2020 at 07:57):
well, I haven't tested that scenario. I'm certainly finding it tough to deal with the variation of meaning of the non-complete code systems, and their idiosyncratic appearance in people's set up
Patrick Werner (May 06 2020 at 08:02):
What about not loading them in the first place? Just ignore alle "not-presen" CS?
Grahame Grieve (May 06 2020 at 08:02):
that's what I did in the first place, but then people expect the validator to not say 'unknown code system' when in fact it is known
Patrick Werner (May 06 2020 at 08:11):
maybe the confusion was that the validator said unknown CodeSystem and throw an error for this?
Patrick Werner (Jan 22 2021 at 11:33):
i have to come back to this. This is still a problem for us as many projects are using not-present CodeSystems as a placeholder.
Patrick Werner (Jan 24 2021 at 10:58):
The german „KBV“ is also struggling a lot with this.
Patrick Werner (Jan 24 2021 at 10:59):
@Maximilian Reith fyi
Maximilian Reith (Jan 25 2021 at 08:35):
yes of course: A lot of codesystems are not present right now, but they are prepared for future usage. Codes are set into a valueset only as an interim solution (compose and expansion), but the validator is giving hard errors, since they are not in the underlying codesystem. We (KBV and Hl7 Germany) would like to change this behaviour as Patrick mentioned. At least an option in the validator to avoid these errors would be necessary
Grahame Grieve (Feb 04 2021 at 23:25):
well, @Patrick Werner, would you like to prepare a test case for the validator that illustrates the issue?
Last updated: Apr 12 2022 at 19:14 UTC
