Creates a new dataset with the given name, description, and schema. The schema provided should be an exact representation of the dataset you are loading. All schema properties must be present in the dataset file you are loading. Data is matched to columns by name, not by order, and headers must match schema columns.