The data mart stores historical data but the operational data store only the latest snapshot. Therefore, even though the historical fact data in the data mart is linked to data in the operational data store, there is a mismatch between the two because the data in the operational data store is updated with the latest snapshot.
You can create lookups to track the relationship between fact data and the data in the operational data store data that contributes to the fact. The extract, transform, and load (ETL) builds calculate this relationship and deliver the historical data to the historical lookup tables in the data warehouse.
A lookup consists of a single data stream, where the objects in the data stream have attributes. The available attributes are determined by associating a template with the lookup. A lookup must include at least an ID attribute to identify each object.