apache/superset · error · DatabaseNotFoundError
Database not found.
Error message
Database not found.
What it means
DatabaseNotFoundError raised in DatabaseUpdateCommand.run() when the model id to update no longer resolves. The whole update runs in a transaction whose error handler reraises wrapped failures as DatabaseUpdateFailedError, but a missing row aborts immediately with not-found.
Source
Thrown at superset/commands/database/update.py:59
from superset.utils.decorators import on_error, transaction
logger = logging.getLogger(__name__)
class UpdateDatabaseCommand(BaseCommand):
_model: Database | None
def __init__(self, model_id: int, data: dict[str, Any]):
self._properties = data.copy()
self._model_id = model_id
self._model: Database | None = None
@transaction(on_error=partial(on_error, reraise=DatabaseUpdateFailedError))
def run(self) -> Model:
self._model = DatabaseDAO.find_by_id(self._model_id)
if not self._model:
raise DatabaseNotFoundError()
self.validate()
if "masked_encrypted_extra" in self._properties:
# unmask ``encrypted_extra``
self._properties["encrypted_extra"] = (
self._model.db_engine_spec.unmask_encrypted_extra(
self._model.encrypted_extra,
self._properties.pop("masked_encrypted_extra"),
)
)
# Depending on the changes to the OAuth2 configuration we may need to purge
# existing personal tokens.
self._handle_oauth2()
# Some DBs require running a query to get the default catalog.
# In these cases, if the current connection is broken thenView on GitHub (pinned to f4587218dd)
Solutions
- Reload the connection list and re-apply the edit to the current id
- Check for concurrent deletion before retrying
- Use the UI's refreshed state rather than a stale dialog
Defensive patterns
Strategy: validation
Validate before calling
if not DatabaseDAO.find_by_id(db_id):
raise LookupError(f"database {db_id} was deleted; reload and retry") Try / catch
try:
DatabaseUpdateCommand(db_id, data).run()
except DatabaseNotFoundError:
# reload connection, re-apply edit to current id
... Prevention
- Shorten edit-dialog lifetime; coordinate concurrent admin edits
- Recheck existence right before PUT
When it happens
Trigger: PUT /api/v1/database/<id> where the connection was deleted concurrently; optimistic UI editing on a stale id.
Common situations: Two admins editing/deleting the same connection; long-open edit dialogs; API retries after a deletion.
Related errors
- Annotation not found.
- Annotation layer not found.
- Annotation layer not found.
- Chart not found.
- Dashboard not found.
AI-assisted analysis of apache/superset@f4587218dd (2026-08-14).
Data as JSON: /api/errors/205e568e668be0f9.
Report an issue: GitHub.