Skip to main content

CREATE DATASET

CREATE DATASET creates a new dataset from a source in the current connection.

Syntax

CREATE [ OR REPLACE ] DATASET [ IF NOT EXISTS ] dataset_name 
{ FROM { table_name | 'scheme://bucket/path/to/file' } | AS query }

Dataset name

A unique name for the dataset in Predibase.

Connection name

A dataset is created from tables or files in the current connection.

FUTURE: To change to a different connection name you can issue the following command:

USE CONNECTION connection_name

Examples

Database table

If the source connection is a database, then the object name will reference a specific table eg:

CREATE DATASET titanic FROM my_table

Object store path

If the source connection is an object store, then specify the path as a string literal eg:

CREATE DATASET titanic FROM 's3://bucket/path/to/titanic.csv'

Database query

FUTURE: Or you can specific query against a specific source:

CREATE DATASET mysql_dataset AS SELECT * FROM mysql_connection.my_table