Background:

During the auto-detection, Power BI:

• Looks for matching or similar column names across tables.
• Checks data types to confirm compatibility.
• Attempts to infer relationship direction, cardinality, and filter propagation.

SQLite database structure meets the first two criteria.

• However, relationship direction and cardinality depend on the exported data. If one input, output or statistic is exported Power BI sets Cardinality to One to one (1:1) and Cross filter direction to Both.

In the example below, the user selected only one input to export in the database. Key columns of the InputsStatistics and Statistics tables match only once so Power BI automatically sets Cardinality to One to one (1:1) and Cross filter direction to Both:

This could be a problem when reusing the dashboard with an updated database which includes more inputs, outputs or statistics.

Also, Power BI could have set other relationships to be Inactive.

Solution 

You can use the Manage relationships dialog to manually create or edit relationships.

The correct number of relationships should be 14; all of them with Cardinality Many to one (*:1), Cross filter direction to Single and Status to Active.

Best Practices

The recommendations to avoid this problem are:

1. Review automatic relationships after connecting to any @RISK SQLite database file.
2. Manually edit relationships right after data is loaded using the Manage relationships dialog box.
3. Include at least two inputs, outputs and statistics in the exported database.

 

Related Article:

Create and Manage Relationships in Power BI Desktop - Power BI | Microsoft Learn