Help Center/ ModelArts/ Data Preparation/ Data Connection/ Managing AI DataLake Data Sources
Updated on 2026-11-06 GMT+08:00
View PDF
Share

Managing AI DataLake Data Sources

When creating a data connection, you need to select a data source. If your data is stored in AI DataLake, you must first create an AI DataLake data source to establish a connection channel between ModelArts and AI DataLake, and then select this data source when creating the data connection. ModelArts only saves the connection configuration and does not store the raw data, which remains stored in the original data source.

Constraints

  • Once a data source is created, it cannot be modified.
  • If AI DataLake is not deployed in your region or the version does not match, you cannot create an AI DataLake data source.

Prerequisites

  • LakeFormation metadata has been created.
  • You have configured a custom policy with the permissions required by the actions on which AI DataLake depends. For details, see the following JSON view configuration. For detailed operations, see Creating a Custom Policy in JSON View. 
    {
    "Version": "1.1",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "lakeformation:catalog:describe",
                "lakeformation:database:describe",
                "lakeformation:dataset:describe",
                "lakeformation:instance:describe",
                "lakeformation:agency:describe",
                "lakeformation:agreement:describe"
            ]
        }
    ]
}

Creating an AI DataLake Data Source

  1. Log in to the ModelArts console.
  2. In the left navigation pane of the console, choose Data Preparation > Data Connections.
  1. In the upper right corner of the Data Connections page, click Connection Configuration.

  2. In the upper right corner of the Data Sources page, click Create.
  3. On the Create Data Source page, configure the required information and click Create Now.
    Table 1 Parameters for creating a data source

    Parameter

    Description

    Example Value

    Data Source Type

    Only AI DataLake is supported.

    AI DataLake

    Name

    Custom data source name. Only letters, digits, hyphens (-), and underscores (_) are supported. Enter 1 to 128 characters.

    data-connect-20260813205323

    Description

    Custom description of the data source. It can contain a maximum of 512 characters.

    -

    Region

    Select a region as needed.

    CN North-Ulanqab1

    Instance

    Select a created AI DataLake instance.

    lakeformation

    Catalog

    Select the created data catalog.

    test

    Database

    Select a database in the selected catalog.

    default
  4. On the Data Sources page, view the created data source.
    Figure 1 Data Sources

Deleting an AI DataLake Data Source

You can delete a data source that is no longer used. Once deleted, it cannot be recovered.

Only the data source record is deleted; the stored files are not deleted. To delete the file, go to the storage container.

  1. Log in to the ModelArts console.
  2. In the left navigation pane of the console, choose Data Preparation > Data Connections.
  1. In the upper right corner of the Data Connections page, click Connection Configuration.
  2. In the Operation column of the target data source on the Data Sources page, click Delete.
  3. In the displayed dialog box, enter DELETE and click OK.
Parent Topic: Data Connection