Why do Queries, used to load Links often use DISTINCT? - OptimalBI/optimal-data-engine-mssql GitHub Wiki

A Link defines the fact that 2 or more Hubs are related.

As such it can and should only hold a single instance of each relationship (there may be many related Satellite instances, but Links relate Hubs).

If the Grain of the Data Source, being used to identify the fact that a Link exists, contains multiple occurrences of the Link, then we need to cut it down to single instances to match the Link requirement.

"Distinct" is one way of achieving this.