apache/superset · error · DatasourceTypeNotSupportedError

DAO datasource query source type is not supported

Error message

DAO datasource query source type is not supported

What it means

Raised by DatasourceDAO.get_datasource when the requested datasource_type is not a key in the DAO's `sources` registry, which only maps DatasourceType.SL (datasets), QUERY, SAVEDQUERY, and SEMANTIC_VIEW to ORM models. It is a DAOException with HTTP status 422. The method resolves a logical datasource type plus a database-side identifier to a concrete model instance, so an unregistered type cannot be resolved at all.

Source

Thrown at superset/daos/datasource.py:64

    return value.replace("\\", "\\\\").replace("%", "\\%").replace("_", "\\_")


class DatasourceDAO(BaseDAO[Datasource]):
    sources: dict[Union[DatasourceType, str], type[Datasource]] = {
        DatasourceType.TABLE: SqlaTable,
        DatasourceType.QUERY: Query,
        DatasourceType.SAVEDQUERY: SavedQuery,
        DatasourceType.SEMANTIC_VIEW: SemanticView,
    }

    @classmethod
    def get_datasource(
        cls,
        datasource_type: Union[DatasourceType, str],
        database_id_or_uuid: int | str,
    ) -> Datasource:
        if datasource_type not in cls.sources:
            raise DatasourceTypeNotSupportedError()

        model = cls.sources[datasource_type]

        if str(database_id_or_uuid).isdigit():
            filter = model.id == int(database_id_or_uuid)
        else:
            try:
                uuid.UUID(str(database_id_or_uuid))  # uuid validation
                filter = model.uuid == database_id_or_uuid
            except ValueError as err:
                logger.warning(
                    "database_id_or_uuid %s isn't valid uuid", database_id_or_uuid
                )
                raise DatasourceValueIsIncorrect() from err

        datasource = (
            db.session.query(cls.sources[datasource_type]).filter(filter).one_or_none()
        )

View on GitHub (pinned to f4587218dd)

Solutions

1. Pass one of the registered types: 'SL' for datasets, 'QUERY' for query results, 'SAVEDQUERY' for saved queries, or 'SEMANTIC_VIEW' for semantic views.
2. If you control the caller, log datasource_type before the call and diff it against DatasourceDAO.sources keys to catch typos or stale values.
3. If you are adding a custom datasource type, register its ORM model by extending the `sources` classvar mapping on DatasourceDAO rather than calling get_datasource with an unknown type.
4. On version upgrades, grep the release notes / UPDATING.md for removed datasource types and migrate stored references.

Example fix

// before
DatasourceDAO.get_datasource('druid', database_id)  # raises DatasourceTypeNotSupportedError (422)

// after
DatasourceDAO.get_datasource(DatasourceType.SL, database_id)  # registered in cls.sources

from superset.daos.datasource import DatasourceDAO
from superset.models.core import DatasourceType  # or the enum's home module

def is_supported_datasource_type(ds_type: str) -> bool:
    return ds_type in DatasourceDAO.sources

def is_known_datasource_type(t: str) -> TypeGuard[str]:
    from superset.daos.datasource import DatasourceDAO
    return t in DatasourceDAO.sources

from superset.daos.exceptions import DatasourceTypeNotSupportedError
try:
    ds = DatasourceDAO.get_datasource(ds_type, db_ref)
except DatasourceTypeNotSupportedError:
    # 422: caller sent an unregistered type; do not retry with same input
    return bad_request(f'unsupported datasource type: {ds_type}')

When it happens

Trigger: Calling DatasourceDAO.get_datasource('SOME_TABLE_TYPE', db_id) with a type not in {SL, QUERY, SAVEDQUERY, SEMANTIC_VIEW}; passing a legacy or custom datasource type string (e.g. 'druid' on builds where it is not registered); sending a typo'd type query parameter through an API that forwards it verbatim to the DAO.

Common situations: Upgrades that rename or remove datasource types (e.g. Druid removal) while old clients or cached dashboards still send the old type; third-party plugins extending the datasource API surface without registering their model in cls.sources; frontend code constructing the type string from a stale enum.

Related errors

• Datasource value is neither id or uuid
• Invalid order column: {order_column}
• created_by_fk_or_editor only supports 'eq'; got '{c.opr}'
• Datasource does not exist
• Datasource does not exist

AI-assisted analysis of apache/superset@f4587218dd (2026-08-14). Data as JSON: /api/errors/f7639acb22fc624d. Report an issue: GitHub.