FHIR Chat · FHIRPath Patch · implementers

Stream: implementers

Topic: FHIRPath Patch

Grahame Grieve (Oct 15 2018 at 20:41):

@Christiaan Knaap FHIR Path Patch tests are in the build at \tests\patch\fhir-path-tests. @Lloyd McKenzie We have 2 weeks to iterate on this and come up with any new rules that we have to state in the text

Grahame Grieve (Oct 15 2018 at 20:42):

this is in regard to GF#18815

Brian Postlethwaite (Oct 15 2018 at 21:57):

I'm working on an implementation of this too.

Brian Postlethwaite (Oct 15 2018 at 21:58):

I'll try take a look through those tests.

Brian Postlethwaite (Oct 15 2018 at 21:58):

Looks like the complete set of issues I had too.

Brian Postlethwaite (Oct 15 2018 at 21:59):

(Will add notes on what assumptions I made)

Christiaan Knaap (Oct 19 2018 at 12:18):

@Grahame Grieve : That is not very long considering a week holiday, but I'll report my findings as I find them. Thanks for the tests, I found them.

nicola (RIO/SS) (Oct 19 2018 at 14:48):

@Grahame Grieve why do we use Parameter resource for FHIRPatch? Isn't it worth to have dedicated resource?! I'm not sure about FHIRPath applicability in context of patching - because "everything is collection". Can we start design from use cases:

• update given for official name
• upsert home phone number

nicola (RIO/SS) (Oct 19 2018 at 14:54):

resourceType: FHIRPatch
operations:
- type: replace
  path: "Patient.name.where(use='official').given.single().materialize()"
  value:  'John'

Can we reuse JSON-patch specification for that - just by replace json path with fhirpath?

Jenni Syed (Oct 19 2018 at 14:56):

@nicola (RIO/SS) how would you identify the specific instance of a resource? How would you expect a batch or transaction to look that has a mixture of creates, patches, etc?

Jenni Syed (Oct 19 2018 at 14:57):

From what I understand, the primary reason we created a FHIRPatch is for the transaction and batch use case (and so we could give a fhirpath option to those that don't want to use json patch or xml patch syntax)

nicola (RIO/SS) (Oct 19 2018 at 14:57):

I understand we need format agnostic patch - but fhirpath is too broad for transformation

Jenni Syed (Oct 19 2018 at 14:58):

note that the json and xml options are also still there

Grahame Grieve (Oct 19 2018 at 20:34):

I'm not sure about FHIRPath applicability in context of patching - because "everything is collection"

I didn't follow this. I am assuming that you think that creating a new resource will do something more than just make a more convenient syntax?

Grahame Grieve (Oct 19 2018 at 20:34):

as for a new resource.... I'm inherently suspicious of a single use resource like this. It would be a simpler syntax, yes.

Grahame Grieve (Oct 19 2018 at 20:35):

there has occasionally been requests for a kind of smart patch, one that you can run against multiple resources.... perhaps there might be a resource like that?

Christiaan Knaap (Oct 21 2018 at 12:29):

@Grahame Grieve : I have a working implementation and posted the comments I found along the way in a gist at http://bit.ly/FhirPatchComments.

Last updated: Apr 12 2022 at 19:14 UTC