Stream: bulk data
Topic: Includes
Michele Mottini (Aug 31 2020 at 13:59):
We found useful supporting includes in our server, and so we added an _include parameter to the bulk export request that specifies the types to include (we did not bother with the search parameter - just the referenced type)
Michele Mottini (Aug 31 2020 at 14:00):
The use case is a client that wants DiagnosticReports and the related Observations (but not all Observation) etc.
Michele Mottini (Aug 31 2020 at 14:00):
Maybe worth considering for the standard ?
Michele Mottini (Aug 31 2020 at 14:01):
(We also have a special 'None' value that disable the automatic include of Practitioner and Location)
Dan Gottlieb (Aug 31 2020 at 15:51):
@Michele Mottini we recently defined the includeAssociatedData param (with an extensible valueset) to enable clients to specify that sort of information (http://build.fhir.org/ig/HL7/bulk-data/export/index.html#query-parameters).
Dan Gottlieb (Aug 31 2020 at 15:52):
We discussed using _include, but want to support behaviors like including only the most recent provenance resource related to the data being retrieved (or excluding related Practitioner resources :) that don't fit with the FHIR definition of the param.
Dan Gottlieb (Aug 31 2020 at 15:52):
Would building on that work for your use cases?
Michele Mottini (Aug 31 2020 at 16:00):
I saw that but the description says 'pre-defined set of _metadata_ associated with the request.' so I thought it did not apply to including resources
Michele Mottini (Aug 31 2020 at 16:00):
The limitation of using those preset is that you have to add a different one for each combination of resources you want to include
Michele Mottini (Aug 31 2020 at 16:25):
Oh wait .... you can define one preset for each resource you want to include and then specify a list of preset, can't you?
Dan Gottlieb (Aug 31 2020 at 17:05):
Yup, the param accepts a list so clients should be able to combine codes.
Dan Gottlieb (Aug 31 2020 at 17:06):
In terms of wording, while our initial use case was metadata, I've been thinking of it as potentially broader. Maybe we should say something in the spec like "When provided, server with support for the parameter and requested values SHALL return or omit a pre-defined set of FHIR resources associated with the request"?
Michele Mottini (Aug 31 2020 at 17:29):
Yes, that seems much better wording to me
Dan Gottlieb (Aug 31 2020 at 18:31):
Cool - I opened https://github.com/HL7/bulk-data/pull/83 to discuss at the Bulk Data Argonaut call this Thursday. @Michele Mottini , would be great if you're able to attend.
Michele Mottini (Sep 01 2020 at 14:18):
When is the call @Dan Gottlieb ?
Karl M. Davis (Sep 01 2020 at 14:24):
(I was wondering the same as Michele, and don't see a bulk WG listed here: https://confluence.hl7.org/)
Dan Gottlieb (Sep 01 2020 at 14:34):
Thursday at 2pm ET - the Argonaut calendar is at http://2020.argo.run
Michele Mottini (Sep 01 2020 at 14:54):
Thanks
Last updated: Apr 12 2022 at 19:14 UTC
