Stream: implementers
Topic: ETL for on_premise sql data to FHIR Format
HadoopHelp (Feb 17 2020 at 14:55):
Abel Stolz (Feb 17 2020 at 14:58):
Your original post is deleted...
HadoopHelp (Feb 17 2020 at 14:59):
Plz below is :-
https://chat.fhir.org/#narrow/stream/179224-india/topic/ETL.20for.20rest.20data.20from.20On-Premise.20to.20FHIR.20Format/near/188213277
Abel Stolz (Feb 17 2020 at 15:03):
In the comments to the issue your link is referring to, you already got the answer that you should rephrase your question, because it is hard to understand. I think you should take the time and diligence to write a more concise question. Then you probably will get your answers.
HadoopHelp (Feb 17 2020 at 15:08):
@Abel Stolz .
i try to explain one more time.
i have sql tables data for patient and others resources .
and trying to convert into json format as require for FHIR JSON.
and every records i am storing into a file and then further reading all the json file through python.
and then putting all the json file into FHIR server using PUT/POST.
Now problem here is when i try to perform ETL for one resources for patient then we are not able reference another resources.
and that is the problem for me.
HadoopHelp (Feb 17 2020 at 15:13):
@Abel Stolz .
i am also looking for someone who has implemented the same use case.
for :-sql rest data how can we put/POST into FHIR server ?
we have huge data with ON_premise sql db and want load into the FHIR SERVER.
so that purpose i trying to do.
for example this patient(1) belongs to;-
"managingOrganization": {
"reference": "Organization/1",
"display": "ACME Healthcare, Inc"
}
then how can we refer that by using ETL.
Lloyd McKenzie (Feb 17 2020 at 16:13):
Have you looked at transaction? That should allow you to import multiple inter-related resources. You can also look at conditional references
HadoopHelp (Feb 18 2020 at 07:25):
how can we check conditional reference here for all the resources tables?is there any predefined way to check that on.
Lloyd McKenzie (Feb 18 2020 at 14:35):
I'm not understanding the question. Conditional references are handled on a per-record basis. If you wanted to do a query in advance, you could I guess, but you'd still probably want to have a conditional reference in case anything had changed in the interim
nicola (RIO/SS) (Feb 20 2020 at 05:11):
You can use Bulk API to efficiently load data into FHIR server. For huge initial backups we do for aidbox/postgres the "in db conversion " trick in - we load original data as tables and then do insert ... select... with on fly conversion inside database - it is usually couple of orders magnitude faster. All refs checks can be skiped at first migration, you can do it in a bulk mode after import.
Last updated: Apr 12 2022 at 19:14 UTC
