A joined dataset allows you to match and associate fields between two datasets, making it easier to use multiple datasets together based on your business scenarios and thereby create a dataset that meets your data requirements.

Create a joined dataset #

Step 1: Go to the dataset module, click "New" in the upper left corner, and select "New joined dataset"

Step 2: After clicking "New joined dataset", you will enter the dataset creation page, where the list of datasets you have permission to access will be displayed on the left

Step 3: After selecting the required datasets on the left, drag them into the blank area in the middle and select the fields to associate

Step 4: After configuring the fields, click "Save" to create the dataset

3.1 Permissions required to create a joined dataset #

On the model configuration page for creating a joined dataset, only datasets under the current project for which you have view or higher permissions are displayed and available on the left

• You can filter the dataset list by clicking the filter button in the upper right corner of the list
  

3.2 Conditions required for dataset association #

• The storage method of the datasets must be the same (for example, both must be ClickHouse; if one is ClickHouse and the other is Hive, Join is not supported). If you attempt to perform a Join on datasets that do not meet the above conditions, the following prompt will appear on the page:
  

3.3 Supported data types #

• Offline extracted datasets (for example, extracted datasets with Hive or ClickHouse as the data source)
• Real-time extracted datasets (for example, Kafka datasets)
• Direct connection datasets
• Direct connection ClickHouse/ByteHouse datasets (the data tables must come from the same cluster)

3.4 Associated field expressions #

• To associate fields that cannot be directly joined due to inconsistent formats or other reasons, you can select "Expression" when choosing the associated field to process it, reducing the need for secondary processing.

3.5 Permission details #

3.5.1 Resource permissions #

• Users can create associated datasets using datasets for which they have "view or higher permissions".
• Datasets for which the user has "no permissions" or only "read-only" permissions cannot be selected.

Editors of a dataset can only use datasets for which the "dataset owner" or themselves have "view or higher permissions" to modify the associated dataset.

3.5.2 Row and column permissions #

Scenario: After dataset 1 joins dataset 2 to create a logical association and generate dataset X, the rules for row and column permissions on dataset X are as follows:
(1) When a user queries this dataset on the visualization page, resource permissions for dataset X are required. The specific rows and columns that can be queried are determined by the row and column permissions the user has for datasets 1 and 2.
(2) When a user views a chart based on dataset X on a dashboard, in addition to dashboard permissions, row and column permissions for datasets 1 and 2 are required; permissions for dataset X are not required.