php - 优化复杂的 mysql 查询

Question

以下查询在 phpmyadmin 上花费了 0.1313 秒。有什么方法可以优化它以使事情变得更快(比如想在 0.00XX 秒内得到它)？索引已添加到正在进行连接的列中。

SELECT m.id, m.civ, m.prenom, m.nom, m.sexe, m.depart, m.date_entree, m.date_sortie, m.login_userid, m.login_passwd, a.rank_id, r.rank_m, r.rank_f, d.user_id AS depID, c.nom AS cordo, z.rank
FROM `0_member` AS m
LEFT JOIN `0_area` AS a ON ( m.id = a.user_id
AND a.sec_id =2 )
LEFT JOIN `0_rank` AS r ON r.id = a.rank_id
LEFT JOIN `0_depart` AS d ON ( m.depart = d.depart
AND d.user_sec =2 )
LEFT JOIN `0_area` AS z ON ( d.user_id = z.user_id
AND z.sec_id =2 )
LEFT JOIN `0_member` AS c ON d.user_id = c.id
WHERE z.rank = 'mod'
ORDER BY nom

Answer 1

您的查询在“Z”别名表中找到的值为“mod”的最终“WHERE”子句中有一个最终的“WHERE”子句，但您的查询是所有 LEFT JOIN 表明您想要所有成员，而不管可能的匹配项您要加入的右侧 table 。

此外，您通过离开并离开用户 ID 将下游连接到“z”表，然后直接重新连接到“0_area”作为直接在用户 ID 上的 A 表，这与 APPEARS 相同无论如何从离开表到“z”表的链接中找到。

就是说，您的成员加入出发然后前往区域...

我的建议(我可以这样重写查询)是颠倒查询的顺序，将您的 Area 表放在首位，并在“sec_id，rank”上提供可用的索引...我将根据键顺序在哪个类别上首先具有较小的子集列...因此 SEC_ID、RANK 或 RANK、SEC_ID。然后对其他表进行简单的 JOIN(不是 LEFT JOIN)...至少从:

SELECT STRAIGHT_JOIN
      m.id, 
      m.civ, 
      m.prenom, 
      m.nom, 
      m.sexe, 
      m.depart,  
      m.date_entree, 
      m.date_sortie, 
      m.login_userid, 
      m.login_passwd, 
      a.rank_id, 
      r.rank_m, 
      r.rank_f, 
      d.user_id AS depID, 
      c.nom AS cordo, 
      z.rank
   FROM 
      `0_area` AS z
          JOIN `0_depart` AS d
             on z.user_id = d.user_id
             and d.user_sec = 2
             JOIN `0_member` AS m
                on d.depart = m.depart
                AND z.user_id = m.id
          LEFT JOIN `0_rank` AS r
             on z.rank_id = .rid
   WHERE
          z.sec_id = 2
      AND z.rank = 'mod'
   ORDER BY
      nom

在您的原始查询中，您有一个连接来自

member
   Links to Area (on member's user ID just to ensure the "sec_id = 2")

由于新查询完全以“area”表作为“Z”别名开始，并且 where 子句明确为“sec_id = 2”值，因此您永远不需要再次反向链接...

Area (only SECID = 2 and rank = mod)
  Links to Depart (on the User's ID)
      Links to Members by (on the depart ID)