LLVM中的可变参数(Varargs in LLVM)

 如果没有正在运行的测试系统，找到适合您问题的解决方案并不容易。 
 
 我发现可疑的是，第一个查询的每个结果行都会调用第二个方法，而且似乎每次打开和关闭数据库连接。 （你的ConnectionHandler可能会实现正确的连接池，但我无法在不知道代码的情况下判断它。如果每次都真正打开和关闭连接，这是你代码中最昂贵的事情）。 
 
 但即使使用正确的连接池，第二种方法也会为每次调用创建并执行新的SQL语句。 这称为N + 1选择查询问题 。 
 
 所以我建议进行以下改进： 
 
 
 
 尝试将两个SQL语句合并为一个（可能使用嵌套的SELECT等）。 那么你不能执行1 + N个查询，这将大大减少执行时间。 
 
 
 如果这不可能，请至少使用PreparedStatement作为第二种方法。 确保在进入循环之前只创建一次预准备语句。 所以你必须在第一种方法中创建和准备它。 因此，您需要第二个数据库连接。 循环之后也分别打开和关闭它。 
 
 
 另一种策略：运行第一个查询，迭代结果集并将所有regCodes放入列表中。 关闭查询后， 通过PreparedStatement使用批处理来处理所有regCodes 。 另一个优点：因此只需要一个数据库连接。 
 
 
 顺便说一下：像Hibernate这样的成熟对象关系映射框架已经提供了开箱即用的策略，以避免或至少减轻1 + N选择问题。 

It is not easy to find a proper solution to your problem without having a running test system to play around.
 
What I find suspicious is the fact that your 2nd method is called for every single result row of your first query, and it seems to open and close database connections every single time. (Your ConnectionHandler might implement proper connection pooling, but I cannot judge this without knowing the code. If the connection is really opened and closed every time, this is by far the most expensive thing in your code).
 
But even with proper connection pooling, your 2nd method creates and executes a new SQL statement for every call. This is called the N+1 Select Query Issue.
 
So I suggest the following improvements:
 
 
 
Try to merge both SQL statements into a single one (maybe using nested SELECTs, etc.). Then you must not execute 1+N queries, which will greatly reduce execution time.
 
 
If that's not possible, use at least a PreparedStatement for the 2nd method. Make sure to create the prepared statement just once before entering the loop. So you'll have to create and prepare it in the first method. Therefore, you need a second database connection. Open and close it before respectively after the loop too.
 
 
Another strategy: Run the first query, iterate the result set and put all regCodes in a list. After closing the query, use batch-processing via PreparedStatement to work off all the regCodes. Another advantage: Therefore only one database connection is required.
 
 
Btw.: Mature object-relational mapping frameworks like Hibernate already provide these strategies out of the box to avoid or at least extenuate the 1+N select problem.