fhir-data-pipes icon indicating copy to clipboard operation
fhir-data-pipes copied to clipboard

Document current parquet schema for the sql-on-FHIR v1 and come up with a plan to standardize it with the community

Open chandrashekar-s opened this issue 1 year ago • 2 comments
trafficstars

Currently the parquet schema that is implemented by fhir-data-pipes and the Pathing for extensions is different. Discuss in the FHIR community and come to consensus on a common schema.

chandrashekar-s avatar Apr 18 '24 14:04 chandrashekar-s

One update here is that there has been some discussions in the community for this standardization and here is a draft document in which various issues are being discussed.

One issue that I recently had to deal with was again around the way that we (i.e., Bunsen) deal with extensions. I had a Parquet file that was generated from an R4 Condition resource but without any extra extension StructureDefinitions. Then I was reading it in another pipeline which had --structureDefinitionsPath set to r4-us-core-definitions. This was causing a strange ArrayIndexOutOfBoundsException where the root cause was the extra assertedDate in the US-Core version which was expected but not present in the input Parquet file.

bashir2 avatar Mar 28 '25 19:03 bashir2

We will experiment with the new proposed schema which is extension independent.

bashir2 avatar Apr 15 '25 14:04 bashir2