Oracle, PostgreSQL and Databricks in Power BI
Authored dataset, field and metric synonyms are included in the generated model's native English (en-US) linguistic metadata. Each synonym remains a separate term, including names with semicolons or quotation marks. Synonyms follow the published table and member identities, including separate roles of the same source table.
Use POWER_BI_TMDL to generate a Power BI project from these source databases.
The project uses DirectQuery. Supply verified source-column metadata when invoking
the compiler, then open the generated Model.pbip and configure credentials in
Power BI Desktop.
| Database | Connection information |
|---|---|
| Oracle | Oracle service/TNS alias and schema. Keep database account credentials outside the catalog. |
| PostgreSQL | Server, optional port, database and schema. |
| Databricks | Workspace hostname, SQL warehouse HTTP path, catalog and schema. |
Dataset mappings can override the schema. For PostgreSQL and Databricks, a dataset catalog override selects its database or catalog. Publication deployment settings do not change the physical source locations.
Table and view sources support fields, measures and the existing Power BI single-column many-to-one relationships. Related-field calculations can use generated SQL projections. Dataset SQL queries retain the authored source SQL; fully qualify table names when the database session defaults are different. PostgreSQL calculations requiring a join between different databases are rejected.
Oracle and Databricks are additions to the existing PostgreSQL, Snowflake and DuckDB source routes. Their connectors do not imply support for every Tableau or Power BI operation. Composite relationships and advanced source projections retain the Power BI target's existing restrictions; unsupported shapes produce diagnostics.
Generation tests validate the source settings, project structure, calculations, relationship mappings and DirectQuery partitions. Opening the project, driver and gateway configuration, authentication and query folding must still be verified in your Power BI environment. There is no automatic switch to Import mode.
For connector requirements, consult Microsoft's Oracle, PostgreSQL, and Databricks documentation.