Create a dataset

Programatically create datasets using the SDK

To create a new dataset, use the create_dataset() method:

from darwin.client import Client

team_slug = "team_slug"
client = Client.local(team_slug="team_slug")

dataset_slug = "your_dataset_name"
dataset = client.create_dataset(dataset_slug)

You'll be able to refresh the Datasets page in V7, and see if your new dataset has been created.

If the dataset you're trying to create has the same name as an existing dataset in V7, you can import files to that dataset instead using the following logic with the get_remote_dataset() method:

from darwin.client import Client
import darwin

team_slug = "team_slug"
client = Client.local(team_slug="team_slug")

dataset_slug = "your_dataset_name"
dataset_identifier = f"{client.default_team}/{dataset_slug}"

try:
    dataset = client.create_dataset(dataset_slug)
except darwin.exceptions.NameTaken:
    dataset = client.get_remote_dataset(dataset_identifier)