This article showcases how you can connect to a JDBC-compliant database using the DataRobot AI Catalog. More specifically, we connect DataRobot to a Snowflake database and load the contents of a specific table.
Creating a new data connection
To create a new data connection, navigate to the AI Catalog and select Add to Catalog > New Data Connection.
Figure 1. DataRobot landing page
Figure 2. AI Catalog
Now you can define a new connection. For this particular tutorial, we need to use Snowflake JDBC Driver 3.5.4 and the URL that points to the database of our interest.
Figure 3. Available data stores
Figure 4. Defining New Connection
After that is done, DataRobot will ask for the Snowflake credentials as seen in Figure 5. These will be used to verify who you are so that you can receive access to the database. DataRobot will also securely save your credentials (if you choose the required option) which means that you will not have to re-insert your data every time.
As far as file and size limits are concerned, these are the same as with the other ingest data methods and depend on the sizing and type of your DataRobot environment. (If you are a DataRobot customer, you can find the latest file size requirements here.)
Figure 5. Defining Credentials
Finally, the dataset is published in DataRobot and we can now start a DataRobot project with it.Figure 6. Published Dataset
Not using Managed AI Cloud?
If you are using DataRobot On-Premise AI Cluster, Private AI Cloud, or Hybrid AI Cloud, your administrator will need to install the JDBC drivers you want to use.
If you’re a licensed DataRobot customer:
Search the in-app documentation for Data connections.