Hi,
Has anyone successfully run OHDSI tooling (e.g., Achilles, CohortMethod) against UK Biobank’s OMOP CDM data on the Research Analysis Platform (RAP)?
As far as I understand, the OMOP tables on RAP are exposed via a Spark-backed database (dxdata/sparklyr, dx xtract_dataset), not a standard RDBMS. DatabaseConnector lists Spark as a supported dbms via JDBC, but I haven’t found confirmation that RAP’s Spark layer exposes an endpoint that external tools can connect to.
So, has anyone gotten DatabaseConnector to connect directly to RAP’s Spark backend? If so, how did you configure the connection? If not, is extract-and-load into a self-hosted Postgres/SQL Server (e.g., via Cloud Workstation) a reasonable workaround (with relation to cost, time), or is there a better-known path?
Thanks