apache/superset · error · SupersetErrorException

The database was not found.

Error message

The database was not found.

What it means

SupersetErrorException with DATABASE_NOT_FOUND_ERROR (HTTP 404, 'The database was not found.') is raised inside the dataset duplicate command's run() when db.session.get(Database, database_id) returns None — the base dataset's stored database_id no longer matches any Database row. The base dataset itself passed validation, so this indicates its connection was deleted (or the metadata is inconsistent) between validate() and run().

Solutions

1. Check the base dataset's database reference (GET /api/v1/dataset/<base_model_id> shows database id) and verify GET /api/v1/database/<id> resolves.
2. If the database was deleted, recreate the connection (same engine/URI) so the orphaned dataset becomes usable, then retry the duplicate.
3. Clean up or re-point orphaned datasets before duplicating them; retrying immediately will keep failing with 404.

# Ensure the base dataset's database still exists before duplicating
from superset.models.core import Database
from superset import db

def base_database_exists(base_model) -> bool:
    return (
        base_model is not None
        and db.session.get(Database, base_model.database_id) is not None
    )

from superset.errors import SupersetErrorException
try:
    DuplicateDatasetCommand(base_id, properties).run()
except SupersetErrorException as ex:
    if ex.error.error_type.name == "DATABASE_NOT_FOUND_ERROR":
        # orphaned base dataset: prompt admin to recreate the connection
        prompt_recreate_database(base_dataset.database_name)
    else:
        raise

When it happens

Trigger: POST /api/v1/dataset/duplicate with base_model_id whose dataset references a deleted database connection; the database row removed concurrently while the duplicate request executes; metadata restored from a backup where datasets outlived their databases.

Common situations: Admin deletes a database connection without cascading cleanup of its datasets; environment migrations that copy datasets but not databases; race between database deletion and an in-flight duplicate.

Related errors

• Database not found.
• Dataset column not found.
• Dataset does not exist
• Dataset parameters are invalid.
• You don't have access to this dataset.

AI-assisted analysis of apache/superset@f4587218dd (2026-08-14). Data as JSON: /api/errors/e91ebb4402fcd0d4. Report an issue: GitHub.

Appendix: source

Thrown at superset/commands/dataset/duplicate.py:70

    @transaction(on_error=partial(on_error, reraise=DatasetDuplicateFailedError))
    def run(self) -> Model:
        self.validate()
        # Declare the high-level avenue before the duplicate touches
        # the session. The change-record listener stamps
        # ``version_transaction.action_kind = 'clone'`` so the new
        # dataset's baseline records read as a clone in the timeline.
        # Method-scoped import — defers the versioning bootstrap path
        # out of this command's module-load graph; see ``changes.py``
        # module docstring for the broader init-order rationale.
        from superset.versioning.changes import ACTION_KIND_CLONE, ACTION_KIND_KEY

        db.session.info[ACTION_KIND_KEY] = ACTION_KIND_CLONE
        database_id = self._base_model.database_id
        table_name = self._properties["table_name"]
        editors = self._properties["editors"]
        database = db.session.get(Database, database_id)
        if not database:
            raise SupersetErrorException(
                SupersetError(
                    message=__("The database was not found."),
                    error_type=SupersetErrorType.DATABASE_NOT_FOUND_ERROR,
                    level=ErrorLevel.ERROR,
                ),
                status=404,
            )
        table = SqlaTable()
        table.override(self._base_model)
        table.table_name = table_name
        table.editors = editors
        table.database = database
        table.is_sqllab_view = True
        if table.sql:
            table.sql = table.sql.strip().strip(";")
        db.session.add(table)

        columns = [

View on GitHub (pinned to f4587218dd)