FHIR Chat · resource ids · implementers

Stream: implementers

Topic: resource ids


view this post on Zulip Brian Postlethwaite (Jan 06 2017 at 20:34):

http://hl7.org/fhir/2017Jan/datatypes.html#id
My server currently limits resource ids to be 64 chars (as per the reference above)
(e.g. [base]/Patient/[Id] where [Id] is restricted according to the ID datatype to 64 chars)
Is this actually a correct limitation? (and hence can be applied to reference types also)

view this post on Zulip Grahame Grieve (Jan 06 2017 at 21:28):

yes

view this post on Zulip Brian Postlethwaite (Jan 07 2017 at 03:08):

While I'm considering ids, when retrieving the _summary=text form of resources, the spec says to return the text, then meta and any mandatory elements. However my server and several others are returning the id field also. Should this be documented?

view this post on Zulip Lloyd McKenzie (Jan 07 2017 at 03:44):

Seems like a pretty important thing to do - seeing as the typical use for summary is to return an overview from which the user can pick one to retrieve the details. Hard to do that (efficiently at least) without the id. Suggest submitting a change request.

view this post on Zulip Brian Postlethwaite (Jan 07 2017 at 04:37):

I was referring specifically to the _summary=text rather than _summary=true

view this post on Zulip Grahame Grieve (Jan 07 2017 at 06:10):

I think the id should always be returned, whatever

view this post on Zulip Richard Ettema (Jan 07 2017 at 16:42):

The current spec defines the behavior of the _summary parameter here: http://hl7.org/fhir/2017Jan/search.html#summary
The '_summary=text' definition is - Return only the "text" element, the 'meta' element, and any mandatory elements
The Resource.id cardinality is defined as 0..1.
Based on the above, my interpretation is that Resource.id is not returned when using '_summary=text'. Sounds like we need a gforge tracker item if the consensus is to return the id in this case.

view this post on Zulip Grahame Grieve (Jan 08 2017 at 06:11):

agree that a gForge tracker item would be a good thing

view this post on Zulip Richard Ettema (Jan 08 2017 at 20:19):

Added #12548

view this post on Zulip Brian Postlethwaite (Jan 11 2017 at 00:46):

http://build.fhir.org/resource.html#id actually explictly defines the Resource Id as up to 64 chars long, and is case sensitive (which I don't handle correctly right now...).
Given the expextation of the case sensitivity of the ids, then a resource reference will also be case sensitive also.
The operation definitions created by the build process output ids like http://hl7.org/fhir/OperationDefinition/structuredefinition-questionnaire, however in the example capabilitystatement (complete) these references are all lower case e.g. ''
http://build.fhir.org/capabilitystatement-base.xml.html

view this post on Zulip Brian Postlethwaite (Jan 11 2017 at 00:46):

Log it to clean up the example complete capability statement to be consistent with what has been generated?

view this post on Zulip Grahame Grieve (Jan 11 2017 at 02:59):

yes

view this post on Zulip Brian Postlethwaite (Jan 11 2017 at 03:13):

Added #12577


Last updated: Apr 12 2022 at 19:14 UTC