Supported platforms
Renart supports the following databases for SQL pipeline assets. Support for a connection does not mean that every database has the same write strategies, catalog features, or data-transfer capabilities.
SQL pipeline assets
Section titled “SQL pipeline assets”| Database | Notes |
|---|---|
| Amazon Athena | Table layout depends on the selected catalog and table format. |
| Amazon Redshift | SQL execution and table/view materialization. |
| Azure Synapse | SQL execution and table/view materialization. |
| ClickHouse | No guided Merge by key mode. Replace and delete/insert require primary-key columns. |
| Databricks | SQL execution and table/view materialization. |
| DuckDB | Embedded database; the local demo needs no warehouse account. |
| Google BigQuery | Supports partitioning and clustering fields for table materialization. |
| Microsoft Fabric | No guided time-interval materialization mode. |
| Microsoft SQL Server | SQL execution and table/view materialization. |
| MotherDuck | Uses the DuckDB SQL workflow with a MotherDuck connection. |
| MySQL | SQL execution and table/view materialization. |
| Oracle | Query-only execution; Renart’s direct SQL runner does not materialize tables or views. |
| PostgreSQL | SQL execution and table/view materialization. |
| Snowflake | Supports clustering fields for table materialization. |
| StarRocks | SQL execution and table/view materialization. |
| Trino | No guided Merge by key mode. Write behavior also depends on the connector and table format. |
| Vertica | SQL execution and table/view materialization. |
The asset’s General → Materialization control shows the modes available for its type and connection. Database permissions, existing objects, and primary keys can impose further requirements. Review Render before executing a new strategy, particularly Full refresh.
These are implemented integrations, not a promise of identical behavior or equal production coverage. Test your database version and materialization strategy against non-critical data during the public alpha.
Files, object storage, and APIs
Section titled “Files, object storage, and APIs”- Load assets copy data between databases, local files, and supported object storage. Configured S3-compatible storage and Google Cloud Storage can be browsed from the source picker.
- HTTP API assets extract JSON records, follow pagination, and load a destination table.
- Notebooks bring source snapshots into local DuckDB. A database source needs a supported typed transfer path; being able to execute SQL against a connection alone is not sufficient.
The connection and asset creation dialogs filter the choices available for the operation. A connection used for storage or source discovery is not necessarily a SQL materialization target.