apache/superset · error · DatasetInvalidError

Dataset parameters are invalid.

Error message

Dataset parameters are invalid.

What it means

DatasetInvalidError (HTTP 422, 'Dataset parameters are invalid.') from the duplicate command aggregates validation failures: base_model_id not found (DatasetNotFoundError appended), base dataset kind != 'virtual' (DatasourceTypeInvalidError — only virtual datasets can be duplicated), a duplicate table_name failing the shared uniqueness check, and editors populated via populate_subject_list failing validation. The response's nested errors list identifies which.

Source

Thrown at superset/commands/dataset/duplicate.py:169

            base_model.database,
            Table(duplicate_name, base_model.schema, base_model.catalog),
        ):
            exceptions.append(DatasetExistsValidationError(table=Table(duplicate_name)))

        try:
            from superset.commands.utils import populate_subject_list

            editors = populate_subject_list(
                self._properties.get("editors"),
                default_to_user=True,
                field_name="editors",
            )
            self._properties["editors"] = editors
        except ValidationError as ex:
            exceptions.append(ex)

        if exceptions:
            raise DatasetInvalidError(exceptions=exceptions)

View on GitHub (pinned to f4587218dd)

Solutions

1. Read the nested errors array: it distinguishes base-model-missing, wrong kind, name collision, and editors problems.
2. Only duplicate virtual datasets (kind == 'virtual'); to copy a physical dataset, create a new dataset over the table instead.
3. Pick a unique table_name scoped to the base dataset's database/catalog/schema; if the message chain mentions a soft-deleted twin, restore or rename.
4. Send editors as a list of existing usernames/roles, or omit the key to default to the caller.

Example fix

// before
POST /api/v1/dataset/duplicate
{"base_model_id": 7, "table_name": "sales_copy", "editors": [{"username": "ghost_user"}]}

// after
{"base_model_id": 7, "table_name": "sales_copy_2026_08", "editors": [{"username": "alice"}]}

# Pre-validate a duplicate request
from superset.daos.dataset import DatasetDAO

def duplicate_request_valid(base_model, new_name: str) -> list[str]:
    problems = []
    if base_model is None:
        problems.append("base dataset not found")
    elif base_model.kind != "virtual":
        problems.append("only virtual datasets can be duplicated")
    if base_model is not None and not DatasetDAO.validate_uniqueness(
        base_model.database,
        __import__("superset.connectors.sqla.models", fromlist=["Table"]).Table(
            new_name, base_model.schema, base_model.catalog
        ),
    ):
        problems.append("table_name already in use")
    return problems

from superset.commands.dataset.exceptions import DatasetInvalidError
try:
    DuplicateDatasetCommand(base_id, properties).run()
except DatasetInvalidError as ex:
    # inspect nested errors: kind, uniqueness, editors, base-missing each get their own UX
    for err in ex.normalized_errors():
        route_error_to_form(err["message"])

When it happens

Trigger: POST /api/v1/dataset/duplicate targeting a physical (non-virtual) dataset; choosing a table_name already used by another dataset on the same database/schema/catalog (including a soft-deleted twin); passing editors entries that don't resolve; passing a nonexistent base_model_id.

Common situations: Attempting to duplicate physical datasets expecting a copy of the table; name collisions like 'sales_copy' already taken; editors list containing removed users.

Related errors

• Dataset parameters are invalid.
• The database was not found.
• You don't have access to this dataset.
• Dataset column not found.
• Changing this dataset is forbidden.

AI-assisted analysis of apache/superset@f4587218dd (2026-08-14). Data as JSON: /api/errors/a91570588cd6f5f5. Report an issue: GitHub.