Question

是否有更好（更高性能）的方式加入MAX（id）（不使用subselect）而不是我的实际方式？我们正在处理庞大的数据库，并且正在尝试尽可能地优化

SELECT e.*, r.report_name FROM report r 
LEFT JOIN engineer e ON e.id = r.id
WHERE e.id = (SELECT MAX(id) FROM engineer WHERE process = r.process)

提前致谢！

Answer 1

也许这会对你有所帮助

WHERE e.id = (SELECT top 1 id 
              FROM engineer 
              WHERE process =r.process
              order by id desc)

Answer 2

我认为那更好

SELECT TOP 1 e.*, r.report_name FROM report r 
LEFT JOIN engineer e ON e.id = r.id
ORDER BY e.id desc

Answer 3

另一个选项：预先聚合整个集合，然后JOIN预先聚合整个集合，以过滤其他表格。这将把它视为一个表，而不是为你遇到的每一行执行相同的SELECT;基于集合的逻辑与逐行激动的行。

SELECT 
    e.*, 
    r.report_name 
FROM 
    report r 
     INNER JOIN -- the WHERE clause effectively makes this an INNER JOIN, so changed for readability
    engineer e ON 
        e.id = r.id
     INNER JOIN 
      (
        SELECT Process, MAX(ID) MaxID
        FROM Engineer 
        GROUP BY Process 
      ) eMax ON 
        e.Process = eMax.Process AND
        e.ID = eMax.MaxID

MSSQL：在MAX（）上左连接

3 个答案: