Add Joins to Data Products

Alation Cloud Service Applies to Alation Cloud Service instances of Alation

Data Product Admins can add joins to the data product properties.

Joins define how tables in a data product relate to each other. By adding joins, you specify how tables should be combined, which helps data analysts query the data correctly and understand relationships between record sets. Joins also enrich the data product with structural context, which improves AI-generated responses when users interact with the data product.

To add joins, Open Data Product Settings on the Joins tab or start editing the data product in advanced mode.

Add a New Join

When no joins exist, the Joins tab displays a message indicating that no joins have been added and provides a New Join button.

To add a join:

  1. Click New Join.

  2. In the Title field, enter a name for the join.

    • The title is required.

    • The title must be unique within the data product. If you enter a duplicate title, the Save button remains disabled.

  3. (Optional) Enter a Description. Use this field to explain how analysts should use the join or provide implementation details. Descriptions enrich the catalog and improve shared knowledge.

  4. In the Tables section:

    • Select the first table (left side).

    • Select the join type: Left Join, Right Join, or Inner Join.

    • Select the second table (right side).

    Only tables already added to the data product are available for selection.

  5. In the Conditions section:

    • Select a column from the first table.

    • Select the corresponding column from the second table.

    Table names are displayed above the drop-downs to help you confirm which table each column belongs to. Only columns included in the data product are available for selection.

  6. Review the SQL preview displayed at the bottom of the dialog. The preview shows the exact SQL join statement generated from your configuration.

  7. Click Save. After saving, the join appears in the joins list. A preview panel displays the SQL definition of the selected join.

You can create multiple joins for the same data product by clicking New Join again at the top of the joins table and repeating the configuration process.

Join Recommendations

Join recommendations are AI-generated when you add a new table and select the Generate descriptions when missing checkbox.

When this option is selected:

  • The system generates descriptions for objects that do not already have descriptions.

  • Join recommendations are generated for the added table.

Recommendations are generated per table addition event. If you do not select Generate descriptions when missing, join recommendations are not generated.

If generation completes without errors, recommended joins appear in the data product. You can keep or delete them manually.

Edit a Join

To modify an existing join:

  1. In the joins list, select the join.

  2. In the top-right corner of the join preview panel, click the Edit (pencil) icon.

  3. In the edit dialog, update any of the following:

    • Title

    • Description

    • Tables

    • Join type

    • Conditions

  4. Review the updated SQL preview.

  5. Click Save.

The changes take effect immediately in the data product configuration.

Delete a Join

To remove a join:

  1. Open the join in edit mode by clicking the pencil icon.

  2. In the bottom-left corner of the edit dialog, click Delete Join.

Use this option to remove joins that were created by mistake or are no longer valid.

Understand Cascading Effects

Join configuration is automatically kept consistent with the tables and columns in the data product.

If you delete a table that is referenced by one or more joins, all joins referencing that table are automatically deleted. This prevents joins from referencing non-existent tables.

If you remove a column that is used in a join condition, the join referencing that column is automatically removed. When editing table columns, a warning dialog indicates that removing columns may affect existing joins.

To prevent schema conflicts, you cannot add a table that is already included in the data product through the New Table action. If you need to change the schema of an existing table, update its columns directly in the table configuration. This ensures joins remain consistent with the table structure.

By defining joins and understanding their cascading behavior, you maintain a consistent, accurate data product configuration that supports analysts and AI-generated insights.