describe_dataset

describe_dataset tells you what’s in a dataset without pulling it into context — column names and types, the row count, and a few sample rows.

Use it before writing a query_datasets SQL statement, so the column names are right first time.

Inputs

Parameter	Type	Required	Description
dataset_id	string	Yes	The ds_... identifier returned by a previous tool.
sample_rows	integer	No	Preview rows. Default 5, max 20.
target_tenant_id	string	No	Super tenants only.

Returns

JSON with the dataset’s schema, row_count, and the head sample.

Notes & limits

• Anything beyond 20 rows should flow through query_datasets with an explicit LIMIT.