aip_sdk.load_dataset

aip_sdk.load_dataset(reference: str | None = None, *, id: str | None = None, client: APIClient | None = None, workspace_id: str | None = None, all_workspaces: bool = False) -> Dataset

Load a dataset by name or by UUID via id=.

A name search reads your session's workspace unless you name one, and raises if none is set; pass all_workspaces=True to search every workspace you can access, where a name used twice resolves arbitrarily. Lookup by id is unaffected.

Parameters

• reference str | None: Dataset name. Always treated as a name search — UUID lookup requires id=. The "name@vN" syntax is not supported: version pinning is not yet implemented (use ds.versions() to inspect available versions).
• id str | None: Dataset UUID for direct GET /datasets/{id} lookup (keyword-only).
• client APIClient | None: Optional API client.
• workspace_id str | None: Search only this workspace.
• all_workspaces bool: Search every workspace you can access.

Returns

• Dataset: Dataset object.

Raises

• DatasetNotFoundError: If the dataset does not exist.
• NotImplementedError: If a version reference ("name@vN") is passed.
• InvalidArgumentError: If neither reference nor id is provided.
• AuthError: If credentials are missing or invalid.
• ForbiddenError: If workspace_id names a workspace you cannot access.
• NoWorkspaceSelectedError: No workspace was passed, none is configured for the session, and all_workspaces was not set.

Examples:

ds = aip.load_dataset("compliance-test-set")  # by name
ds = aip.load_dataset(id="uuid-...")  # by UUID
df = ds.download()