Using the statistics described, the optimizer chooses the best join strategy. Because Microsoft Jet uses a cost-based optimizer, each possible join combination is examined to determine which will yield the least costly query execution.
In choosing a join combination, the optimizer first selects a base-table access strategy. The optimizer then stores the estimated number of records returned and a cost — how expensive it is to read the table. The optimizer then generates all combinations of pairs of tables and determines the cost of each join strategy. Finally, the optimizer adds tables to the joins and continues to calculate statistics until the cheapest strategy for the entire query is found.