Stream: implementers
Topic: Resolving versioned references
Nik Klassen (Dec 07 2018 at 20:25):
If I have an Encounter with a subject: Patient/1/_history/1 should I be able to search for this Encounter using subject==Patient/1 for as long as the version of Patient/1 is 1? Should history references be considered first class citizens with all the integrity checks that a local resource would normally get, or can they be considered the same way as a resource on another FHIR server? I see this primarily as an simplification for indexing and constructing compartments. I tried to figure out what HAPI FHIR's behaviour was, but it doesn't seem to handle versioned resource references very well.
Grahame Grieve (Dec 07 2018 at 20:52):
it's generally a difficult area.
Nik Klassen (Dec 07 2018 at 21:00):
I wasn't able to find a section of the spec that talked about this behaviour, other than the part about resolving them in a bundle. Is there any further clarification on how to deal with these references?
Grahame Grieve (Dec 07 2018 at 21:02):
not a lot
Grahame Grieve (Dec 07 2018 at 21:02):
we've argued about it before
Lloyd McKenzie (Dec 11 2018 at 01:58):
@Nik Klassen, feel free to submit a change request to act as a central place for formally arguing about it some more - and perhaps landing on an answer :)
Nik Klassen (Dec 11 2018 at 15:17):
This certainly seems like a challenging problem. I'll see if I can formulate some opinions as I look into it further
Christiaan Knaap (Dec 17 2018 at 09:50):
Last updated: Apr 12 2022 at 19:14 UTC
