How do you optimize data storage and query performance in Snowflake using clustering keys or partitioning?