Overview
The Datasource Reference source type lets you create a new datasource that reads data from another existing datasource. This is the foundation for derived datasources that compute or combine data from multiple sources.
Configuration
| Field | Description |
|---|
id | The ID of an existing worker-mode datasource to use as a data source. |
Only worker-mode datasources (datasources that pull and store data) can be used as references. The referenced datasource must have a live version with data.
Use Cases
Datasource references are used in several special configurations:
| Configuration | Description |
|---|
| Merged Datasources | Combine two datasources by joining them on a common field. See Special Configurations. |
| Best Products | Compute top-performing products by combining a product catalog datasource with an analytics datasource. |
| Top Attributes | Identify preferred product attributes (brands, categories) per user based on behavioral data. |
| Bought Together (Cross-sell) | Identify products frequently purchased together. |
How It Works
- The new datasource reads data from the referenced datasource’s live storage instead of an external source.
- Fields from the referenced datasource are available for mapping in the new datasource.
- When the referenced datasource updates, the dependent datasource can automatically refresh (depending on the configured refresh schedule).
You can chain datasource references to build multi-step data pipelines — for example, merge a product catalog with CRM data, then compute best products on the merged result.
