Skip to main content

Datasets

Datasets are tables/objects that are imported from respective connections which can be used to train models and run queries. Datasets are dynamic, so a change in a dataset from the source will be reflected in Predibase upon refresh.

Attributes

id    ID of the dataset

name     Name of the dataset

author   Creator of the dataset

object_name   Name of object used to create Predibase dataset (S3 object path, file name, etc.)

connection_id     ID of the connection from which the dataset was created

status    Current status of the dataset (connected, refreshing, errored, etc.)

created     Time of dataset creation

updated    Time of last update to dataset

Methods

Dataset.get_connection

Dataset.get_models

Dataset.get_fields