Microsoft SQL Server
Connect a Microsoft SQL Server database to Lakehousecat using a connection URI.
Connection Fields
| Field | Required | Description |
|---|---|---|
| Connection URI | Yes | Full connection string including credentials, host, port, and database name. Example: mssql://user:password@host:1433/mydb |
| Use SSH | No | Enable if the database is behind an SSH bastion host. When enabled, SSH settings fields appear: Host, Port, Username, Password, Private Key, and Private Key Passphrase. See SSH Tunneling. |
Prerequisites
- Admin or Builder role in Lakehousecat
- A SQL Server login with read access to the schemas and tables you want to expose
- Network connectivity from the Lakehousecat backend to the SQL Server instance (direct or via SSH)
- The Microsoft ODBC driver, provisioned once through a job definition (see below)
Provision the ODBC driver (one-time)
Lakehousecat does not ship Microsoft's ODBC Driver 18 for SQL Server. You fetch it once from
Microsoft through the msodbcsql_driver_provisioning job definition and accept Microsoft's licence
terms while doing so. Then load through msodbcsql_load and validate the connection or discover
the schema through msodbcsql_schema_validate. The driver works on amd64 and arm64.
Validate and load calls made directly against the platform API don't work for SQL Server. They return a message that points to these job definitions.
See Vendor Drivers for the steps and network requirements.
Notes
- Use a read-only login to follow the principle of least privilege.
- For large databases, configure schema and table filters in the Filters tab after creation to limit the scope of semantic extraction.
Next Steps
After provisioning the driver, load the datasource through the msodbcsql_load job definition, then open the Operations tab to trigger semantic extraction.