How would you optimize a query that joins a massive transaction table with a smaller dimension table?
