Stream: implementers
Topic: GF#17032 - Mapping Unknowns
Grahame Grieve (Jun 13 2018 at 02:31):
in GF#17032 Lloyd asks for a way to map an unknown element. Well, you can't map an unknown element to nothing in a language that does things. But we could do this:
group for type+types Binary extends Resource input src : BinaryR2 as source input tgt : Binary as target "Binary-contentType" : for src.contentType make tgt.contentType "Binary-content" : for src.content make tgt.content "Binary- securityContext" : for src. securityContext log "SecurityContext is lost" endgroup
Grahame Grieve (Jun 13 2018 at 02:32):
e.g. add an explicit log option to make this stuff clear. You can use log in any statement....
Jean Duteau (Jun 13 2018 at 02:48):
I guess that is an option but I actually disagree with the tracker item. it's a mapping language and we have nothing to map to. You can indicate that via a comment in the mapping file. Why do we need an actual mapping statement that does nothing for this?
Grahame Grieve (Jun 13 2018 at 02:52):
well, I did think about just adding a comment in the mapping. it would work, but I realised, on reflection, that I often had to use FHIRPath log statements slipped inside conditions to figure out what was going on... so why not make it explicit?
Jean Duteau (Jun 13 2018 at 02:53):
i guess that does make it explicit.
Lloyd McKenzie (Jun 13 2018 at 02:56):
I like the proposed solution
Jean Duteau (Jun 13 2018 at 02:57):
i just don't think it does what you think it does but I don't agree with what you want. :)
Jean Duteau (Jun 13 2018 at 02:59):
from my ccda work, i don't think that FHIR mapping files will become the default means of enumerating a mapping. it's equivalent to telling people to look at the schematron to see the ccda conformance statements. so there needs to be less-programming way of expressing mappings and that is where you'd want to express that such and such an element isn't mapped.
Jean Duteau (Jun 13 2018 at 02:59):
but for the existing r2 to r3 mappings and such, the log statement is sort of equivalent to a comment
Grahame Grieve (Jun 13 2018 at 02:59):
the problem is that expression mappings in any less comprehensive way is misleading.
Lloyd McKenzie (Jun 13 2018 at 02:59):
It allows you to be explicit in the mapping definition and it also spits out the mapping at runtime. I agree it's evil to read, but if you want to see the details of what's happening it's where you need to go.
Last updated: Apr 12 2022 at 19:14 UTC
