The output of the FHIR Data Pipes ETL Pipelines is conversion of raw FHIR resources to a Parquet-on-FHIR schema representation. This takes place for each resource type and follows the FHIR Data Pipes Schema Mapping Rules.
The generated columnar Parquet files provide the 'base data warehouse' that can be queried using any parquet-aware tools (e.g a SQL based query engine) or further transformed via the view layer into materialized views.
Go to developer documentation