Dataset creation consists of three parts: creating a new dataset, creating a new associated dataset, and creating a new folder.

2.1 Create a new dataset #

2.1.1 Entry point #

• Click the plus sign to create a dataset.

2.1.2 Select a data source #

(1) Directly connect to an existing data source
Users can create a new dataset from an existing data connection. First, select the target connection from the data connections, then click "Confirm".

• Before version V2.50.0, as shown in the figure below

  

• From version V2.50.0 onwards, as shown in the figure below

(2) Create a new data connection
From version V2.62.0 onwards, an entry point for permission application has been added in the pop-up window for creating a new dataset.
When the user has "Project Edit" permission, the pop-up window for creating a new dataset appears as follows:

When the user does not have "Project Edit" permission, there are usually two situations:
Situation 1: For users whose system has a permission application module, "Project Edit" permission can be enabled by clicking "Notify administrator by system email".
Situation 2: For users whose system does not have a permission application module, please contact the project administrator to request "Project Edit" permission.
After enabling the above "Project Edit" permission, follow the Data Connections Overview instructions to create a new data connection. Once completed, proceed directly to the next step to configure the model.

2.1.3 Dataset configuration details #

• On the model configuration page, first select either Custom SQL or a table from the database on the left, then drag and drop it to the canvas on the right using the mouse.

• You can click "Custom SQL" under the data connection on the left, drag it to the canvas on the right, and then write your own SQL to create the dataset.
  
• You can also click a table on the left, drag it to the canvas on the right, and create the dataset directly from the table.

Users can also make additional configurations. For details, see:

2.1.4 Preview and save #

Click the "Preview" button on the right to preview and confirm the data. Once confirmed, click the "Save" button to save.

• In the pop-up window, set the dataset name and description, select the save path, and starting from V2.50.0, you can configure Create Dataset Usage Instructions(未找到该功能，不发布) during dataset creation.

• After saving, you can view the details of this dataset. You can click the "Visual Query" button in the upper right corner to use this dataset Visual Query Overview, or click "Edit" to proceed to the above Model Configuration step.

Caution: After creating a dataset, you need to wait for a certain period before the dataset can be previewed.

2.2 Create an associated dataset #

An associated dataset allows you to match and associate fields between two datasets, making it convenient to combine multiple datasets based on your business needs and thereby create a dataset that meets your data requirements.
Operation entry:

Detailed description: Joined datasets

2.3 Create a folder #

Click the plus sign to create a folder, and fill in the name, description, and folder address.
Operation entry:

Details to fill in:

2.4 More configurations #

• V2.74.0 and later, after the user uploads an Excel/CSV local file, the system will perform a "file merge" operation to reduce the number of base tables created and alleviate cluster pressure.
• V2.74.0 and later, for redshift and las_jdbc datasets with direct connections, the "Insert Parameter" feature has been added, allowing users to embed relevant parameters in the dataset according to specific requirements.