Stream: IG creation
Topic: how to slice on binding
Brian Reinhold (Nov 22 2018 at 14:01):
What distinguishes my Observation.component elements is the binding. But the FHIR slicing tools do not seem to support it. How can I effectively slice on the binding?
Lloyd McKenzie (Nov 22 2018 at 15:04):
I would have thought you'd want a fixed value for Observation.component.code most times (and would use pattern). If a particular Observation.component.code is allowed to be any of a range of codes, then you'll have to do a binding for all of them (even if it's only a single-code value set) and then slice by value.
Brian Reinhold (Nov 23 2018 at 13:06):
I would have thought you'd want a fixed value for Observation.component.code most times (and would use pattern). If a particular Observation.component.code is allowed to be any of a range of codes, then you'll have to do a binding for all of them (even if it's only a single-code value set) and then slice by value.
@Lloyd McKenzie This appears to be what I need to do (though not sure how to accomplish the above because that is what I thought I was doing before I tried the pattern approach).
I have four different bindings
1. a set where the codes come from a fixed partition of MDC and the code is fixed and system is fixed to MDC 2. a set where the codes come from all other partitions of MDC except the above and the code is different and the system is fixed to MDC 3. a set where the codes come from subset A of a code system and the codes are variable and the system is fixed to ASN1
4. a set where the codes come from subset B of a code system and the codes are variable and the system is fixed to ASN1
I currently DO specify the binding in each slice.
Lloyd McKenzie (Nov 23 2018 at 15:38):
After some offline discussion w/ Brian, we landed on slicing Observation.component by value with a discriminator of "code" with a distinct value set specified for each slice.
Brian Reinhold (Nov 23 2018 at 16:28):
After some offline discussion w/ Brian, we landed on slicing Observation.component by value with a discriminator of "code" with a distinct value set specified for each slice.
@Lloyd McKenzie I am getting messages as follows in the build for my value sets:
[java] No value set at Observation.component.code (url = 'http://hl7.org/fhir/uv/phd/ValueSet/Quantity11073MDC')
which makes sense given that the URL is where the value set will be located when done in the continuous integration or its final resting place. At least that is where I think its going to go. Of course it doesn't exist there now. Is there a special way to define the URLs to the value sets which are currently only in the IG? Is there a relative way and will that be okay when all is said and done?
Lloyd McKenzie (Nov 23 2018 at 17:40):
The value set with that canonical URL needs to be defined in your IG and referenced by your IG
Brian Reinhold (Nov 23 2018 at 17:50):
The value set with that canonical URL needs to be defined in your IG and referenced by your IG
YEa, my bad. Forgot to do that. I was too proud of myself for actually making the value sets and fixing the sliced components only forgetting to add them to the guide.
However I am getting these kinds of errors everywhere
Could not match discriminator (code) for slice Observation.component:relativeTimeComponent in profile http://hl7.org/fhir/uv/phd/StructureDefinition/PhdNumericObservation - does not have fixed value, binding or existence assertions
probably because all my value sets have this error
Quantity11073MDC error Error from server: Unable to provide support for code system urn:iso:std:iso:11073:10101 (see Tx log)
Lloyd McKenzie (Nov 23 2018 at 23:07):
Have you defined that code system?
Lloyd McKenzie (Nov 23 2018 at 23:07):
Also, does the relativeTimeComponent define and Observation.component:relativeTimeComponent.code element that has a value set binding?
Last updated: Apr 12 2022 at 19:14 UTC
