Stream: implementers
Topic: mapping V2 to FHIR
Elkhan Yusubov (Jun 12 2019 at 18:10):
Mapping documents from HL7 v2 to HL7 FHIR?
@Grahame Grieve presented a link: https://docs.google.com/spreadsheets/d/13pgda5xl-PwCgB9j0axyymwwv7RJVcrlzY8Ah1y1Y1M/edit#gid=0
I am getting 404 error from Google... please guide for current doc?
Craig Newman (Jun 12 2019 at 18:52):
i'm not sure what Grahame was presenting, but we have a number of google docs with mappings for various segments and data types. Was there something in particular you were looking for?
Michael O'Keefe (Jun 12 2019 at 18:53):
I'm not sure which one Yusubov was referencing, but I'm curious to see the URL that Grahame posted to the mapping Javascript Sandbox proposal (I wasn't fast enough to copy the URL down before the slide advanced)
Jeffrey Taylor (Jun 12 2019 at 19:40):
The github link he posted for the javascript sandbox proposal is here: https://github.com/FHIR/interversion/tree/master/engine
nicola (RIO/SS) (Jun 12 2019 at 20:55):
@Grahame Grieve have you seen this - https://github.com/HealthSamurai/jute-example
nicola (RIO/SS) (Jun 12 2019 at 20:55):
Here is a parser - https://github.com/HealthSamurai/hl7grok/tree/6217cc972f4ecb90f5edd1dd70f6e0b514e6c494
nicola (RIO/SS) (Jun 12 2019 at 20:57):
As well here is my 4 years old experiments - https://github.com/niquola/hl7.js
Elkhan Yusubov (Jun 13 2019 at 02:57):
@nicola (RIO/SS) @Jeffrey Taylor here is the screen-shot of that google spreadsheet on mapping. Do you know how to get that spreadsheet?
IMG_20190612_104831145_HDR.jpg
Craig Newman (Jun 13 2019 at 13:52):
Data Type mapping spreadsheets are at https://confluence.hl7.org/display/OO/Data+Type+Mapping
Segment mapping spreadsheets are at https://confluence.hl7.org/display/OO/Segment+Mapping
We are still working to finalize a message mapping format, but the initial thoughts are at https://confluence.hl7.org/display/OO/Message+Mapping
Grahame Grieve (Jun 13 2019 at 14:06):
I will think about Jute. but as the FHIR mapping language experience shows, declarative mappings have lots of advantages, but people think iteratively
Vadim Peretokin (Jun 17 2019 at 14:26):
Having done both extensively, declarative mappings are much, much more nicer to work with. Takes some time to figure out all the "patterns" you could do with the declarative kind but then what you write is significantly shorter than using the imperative version
Grahame Grieve (Jun 17 2019 at 17:47):
yes. I agree. but that doesn't mean people will actually do that
Jay Lyle (Jun 18 2019 at 02:06):
Nice preso on the V2-FHIR work and how spreadsheets break. And how FML is better.
An initial look suggests FML wants StructureDef on both sides. How do you do that with V2?
Grahame Grieve (Jun 18 2019 at 03:06):
thanks. there is an implicit one. And there's an explicit one hanging around somewhere. but you don't need it explicit - you just have to write a data provider in code instead
Last updated: Apr 12 2022 at 19:14 UTC
