Supported Connectors
Multi-schema discovery is available for JDBC-based connectors that support schema-level separation. DFS-based datastores (Amazon S3, Azure Datalake Storage, Google Cloud Storage) do not support multi-schema creation because they do not use a catalog/schema hierarchy.
Connector Reference
The table below shows each supported connector, the field used for catalog selection (if any), and the field used as the schema target for multi-select.
| Connector | Catalog Field | Schema Field |
|---|---|---|
| Athena | Catalog | Database |
| BigQuery | Project ID | Dataset ID |
| Databricks | Catalog | Database |
| DB2 | None | Schema |
| Dremio | None | Schema |
| Hive | None | Schema |
| MariaDB | None | Database |
| Microsoft SQL Server | Database | Schema |
| MySQL | None | Database |
| Oracle | None | Schema |
| PostgreSQL | Database | Schema |
| Redshift | Database | Schema |
| SAP HANA | Database | Schema |
| Snowflake | Database | Schema |
| Synapse | Database | Schema |
| Timescale DB | None | Schema |
| Trino | Catalog | Schema |
Note
Fabric Analytics, Presto, and Teradata do not support multi-schema discovery. See the Available Datastore Connectors page for the full list with multi-schema support status.
Understanding the Columns
Catalog Field
The two columns give the label each field carries in the connector's own form, which differs between connectors: what plays the schema role is called Schema on some, Database on others, and Dataset ID on BigQuery.
The Catalog Field is the first-level hierarchy used to group schemas. Where it reads None, no catalog selection is needed: schemas are discovered directly from the connection.
Examples:
- PostgreSQL: The catalog is the
Database. You first select a database, then discover schemas within it. - BigQuery: The catalog is the
Project ID. You first select a project, then discover the datasets within it. - Oracle: There is no catalog level. Schemas are discovered directly.
Schema Field
The Schema Field is the target for multi-select. This is the level at which individual source datastores are created.
Note
On MySQL and MariaDB, the field carrying this role is labeled Database, because these systems have no separate schema concept. Each database is treated as a schema for multi-schema creation. Hive has no database field at all: its schema target is labeled Schema.
Two-Step vs. Single-Step Discovery
| Flow Type | Connectors | Description |
|---|---|---|
| Two-step (Catalog → Schema) | PostgreSQL, Snowflake, BigQuery, SQL Server, Synapse, Databricks, Redshift, Trino, Athena, SAP HANA | First select a catalog, then discover and select schemas within it. |
| Single-step (Schema only) | Oracle, DB2, MySQL, MariaDB, Hive, Dremio, Timescale DB | Schemas are discovered directly without a catalog selection. |