Due to the flexible structures of SQL, no general approach works efficiently for all kinds of queries. Some special kinds of queries can be further optimized for better performance. In this dissertation, we study two kinds of such queries: one is queries having non-aggregate subqueries and the other is queries having redundancy. To deal uniformly with non-aggregate subqueries in SQL, we propose the nested relational approach based on the nested relational model. To deal with redundancy in complex SQL queries, we propose the redundancy awareness method with the introduction of the for-loop operator. The main contribution of this dissertation is to provide more efficient solutions for these two kinds of queries than existing techniques.