Using Pre-Loaded Data

The ExistingTableSource adapter wraps a DuckDB table or view that you've already loaded — through FHIRDataLoader, custom ETL, or any other mechanism — into the standard source adapter API.

Why ExistingTableSource Exists

If you're already using FHIRDataLoader to load FHIR data into DuckDB, your workflow doesn't need to change. ExistingTableSource exists to provide a uniform API so that code written against the adapter interface works identically whether data comes from files, Postgres, CSV, or pre-loaded tables.

Quick Start

import fhir4ds
from fhir4ds.sources import ExistingTableSource

con = fhir4ds.create_connection()

# Assume 'staged_fhir_data' was created via custom ETL
source = ExistingTableSource("staged_fhir_data")
fhir4ds.attach(con, source)

# A 'resources' view is now available for measures/FHIRPath

Schema Requirements

The existing table or view must conform to the standard FHIR4DS schema contract (id, resourceType, resource, patient_ref). For details on the required types, see the Data Sources & Zero-ETL Schema Contract.

FHIRDataLoader vs Adapter Approach

Aspect	FHIRDataLoader Only	With ExistingTableSource
Data loading	loader.load_directory(...)	Same — loader.load_directory(...)
API pattern	Direct table access	attach() / detach() lifecycle
Schema validation	Manual	Automatic at attach() time
Swappable sources	No	Yes — same code works with any adapter
Breaking changes	None	None — additive only

Migration Path

If you're using FHIRDataLoader today, there are no breaking changes. To adopt the adapter pattern:

1. Keep your existing FHIRDataLoader code.
2. Add ExistingTableSource() and fhir4ds.attach() after loading.
3. Downstream code that uses resources works unchanged.

This is a zero-risk migration — ExistingTableSource validates the schema and passes through to the existing table.