Using Computed Fields in Dataset Models

Overview

After creating a Computed Field, you can attach it to a Dataset Model as a node in the model's relationship graph. This allows Sourcesible to join the Computed Field's results with other datasets in the model — making the computed data available for use in Audiences and other downstream features.

The workflow follows this order:

1. Create a Dataset Model (with a Master Dataset and any Related or Event Datasets)
2. Create a Computed Field scoped to that Dataset Model
3. Return to the Dataset Model and attach the Computed Field as a node

2. Prerequisites

Before attaching a Computed Field to a Dataset Model, ensure the following are in place:

• A Dataset Model exists with at least a Master Dataset configured. See Dataset Models setup guide.
• At least one Computed Field has been created and its Status is Active. See Computed Fields setup guide.
• The Computed Field was built using the same Dataset Model you intend to attach it to.

3. Attaching a Computed Field to a Dataset Model

3.1 Open the Dataset Model

1. In the left navigation, click Dataset Models.
2. From the Dataset Models list, click the model you want to update. The model's relationship graph canvas will open.

3.2 Add a Computed Field Node

1. On the canvas, click the + connector button on the Master Dataset node to open the node type menu.
2. From the dropdown, select Create a COMPUTED FIELD — Filter based on calculated fields from each customer.

3. A right-hand panel will open showing Step 1 — Select Computed Field.

3.3 Select the Computed Field (Step 1)

1. In the Computed Fields panel, you will see only the Computed Fields that were created from this Dataset Model. Select the Computed Field you want to attach by clicking its radio button.
2. Click Confirm to proceed to Step 2.

3.4 Configure the Relationship (Step 2)

1. In the Configuration panel, review the auto-populated Name field — this defaults to the Computed Field's name and can be left as-is.
2. Under Relationship, configure the join between the Master Dataset and the Computed Field:
   • The left dropdown shows the Master Dataset name (e.g., students_Azure) — this is fixed.
   • The middle dropdown sets the relationship type — select 1:many for a standard one-to-many join.
   • The right side shows the Computed Field name — this is fixed.
3. In the column mapping row below, select the join key column from the Master Dataset (left dropdown) and the matching column from the Computed Field (right dropdown). These columns must share the same values — typically a customer or user ID (e.g., student_id = student_id).
4. Click Save.

Sourcesible will update the Dataset Model canvas and display a confirmation banner: Your dataset model has been update successfully. The Computed Field will now appear as a COMPUTED FIELD node connected to the Master Dataset in the graph.

4. Test Your Configuration

After saving, verify the Computed Field node appears correctly on the canvas:

• ✅ A COMPUTED FIELD node is visible, connected to the Master Dataset node with a 1:m relationship label
• ✅ The node displays the correct Computed Field name
• ✅ The Dataset Model's overall Status remains Active

5. Next Steps

With the Computed Field attached to your Dataset Model, you can:

• Build an Audience using the Computed Field — Use the computed data as a filter condition when creating Audience segments
• Add more Computed Fields to the same Dataset Model — Repeat the steps above for each additional Computed Field
• View the Dataset Model relationship graph — Inspect how all datasets and Computed Fields are joined within the model

6. Tips and Troubleshooting

Computed Field Not Appearing in the Selector

Symptom: When selecting a Computed Field in Step 1, the list is empty or the expected field is missing.

Cause: Either the Computed Field was created using a different Dataset Model, or its Status is not Active.

Fix:

1. Go to Computed Fields and locate the field you want to use.
2. Confirm that its Status is Active — if it shows as Inactive or Invalid, open the field and resolve any errors before retrying.
3. Confirm that the field was created using the same Dataset Model. If not, create a new Computed Field scoped to the correct model.

Join Key Columns Not Matching — No Data After Activation

Symptom: The Computed Field node is added successfully, but Audiences using this Dataset Model return no results or unexpectedly small counts.

Cause: The join key columns selected in the Relationship configuration do not share the same values across the Master Dataset and the Computed Field.

Fix: Open the Dataset Model and click the COMPUTED FIELD node to inspect its join configuration. Verify that the selected columns contain matching values by running a quick check in your warehouse:

-- Check overlap between Master Dataset and Computed Field on the join key

SELECT COUNT(*)

FROM master_table m

INNER JOIN computed_field_table c

ON m.student_id = c.student_id;

If the count is zero or unexpectedly low, the join key columns may contain different data types or formatting. Align them in the warehouse or update the Computed Field SQL to normalize the key before joining.

Dataset Model Shows "Invalid" Status After Adding Computed Field

Symptom: After saving, the Dataset Model's status changes to Invalid on the Dataset Models list page.

Cause: The Computed Field's underlying SQL query references tables or columns that are no longer accessible, or the Computed Field itself has an error.

Fix:

1. Go to Computed Fields and open the Computed Field you attached.
2. Click Run Preview to check for SQL errors.
3. Fix any errors in the SQL query and save the Computed Field.
4. Return to the Dataset Model — the status should resolve to Active automatically once the Computed Field is healthy.