优化IN子查询

Question

我有一个worker表和一个关联的workerGeofence表。

CREATE TABLE IF NOT EXISTS `workergeofences` (
`ID` int(11) NOT NULL,
  `WorkerID` varchar(20) NOT NULL,
  `GeofenceID` int(11) NOT NULL,
  `isActive` tinyint(4) NOT NULL
) ENGINE=InnoDB AUTO_INCREMENT=107 DEFAULT CHARSET=latin1;

我只需要返回在workerGeofences表中具有至少一个isActive为1的条目的工人。

我可以通过以下方法获得所需的结果：

    SELECT distinct w.ID, Title, FName, SName, Email, Birthday, Address, Phone, description,
 companyID 
FROM Workers w WHERE companyID = ? 
and w.ID IN (SELECT WorkerID FROM WorkerGeofences WHERE isActive <> 0)
    limit ?,10

但是in子查询是详尽无遗的，因为当我运行解释时，我可以看到它正在扫描整个表。 我该如何解决？

Answer 1

您处在正确的轨道上，但是您不需要select distinct 。 除非您知道有重复项，否则这会减慢查询速度-这不太可能，因为您正在选择WOrkers.Id 。

SELECT w.* 
FROM Workers w 
WHERE w.companyID = ? AND
      EXISTS (SELECT 1
              FROM workerGeofences wg
              WHERE w.ID = wg.WorkerID AND wg.isActive <> 0
             )
LIMIT ?, 10;

然后，对于此查询，您需要在Workers(CompanyId, Id)和workerGeofences(WorkerId, isActive)上建立索引。

注意：为方便起见，我仅输入select * 。 我假设所有列都来自Workers表。

Answer 2

首先，您的加入是错误的！ 您没有在两个表上比较任何公共列，则应在其中添加workerGeofences.workerID = w.id，如下所示：

SELECT  w.ID, Title, FName, SName, Email, Birthday, Address, Phone,
        description, companyID
    FROM  Workers w
    join  workerGeofences
    WHERE  workerGeofences.workerID = w.ID companyID = ?
      and  w.ID IN (
        SELECT  WorkerID
            FROM  WorkerGeofences s
            WHERE  isActive <> 0
              and  s.workerID = w.id
                   )
    limit  0,10 

其次，您没有从第二个表中选择任何内容，因此联接是不必要的，并且在您的IN语句中，您没有在比较正确的ID，因此您的查询应为：

SELECT  w.ID, Title, FName, SName, Email, Birthday, Address, Phone,
        description, companyID
    FROM  Workers w
    WHERE  companyID = ?
      and  w.ID IN (
        SELECT  WorkerID
            FROM  WorkerGeofences s
            WHERE  isActive <> 0
              and  s.workerID = w.ID
                   )
    limit  0,10 

另外，您可以为此使用EXISTS（）。

SELECT  w.ID, Title, FName, SName, Email, Birthday, Address, Phone,
        description, companyID
    FROM  Workers w
    WHERE  companyID = ?
      and  exists 
      ( SELECT  1
            FROM  WorkerGeofences s
            WHERE  isActive = 1
              and  s.workerID = w.ID
      )
    limit  0,10

Answer 3

为了确保使用JOIN的完整性：

SELECT DISTINCT w.ID,
       w.Title, 
       w.FName, 
       w.SName, 
       w.Email, 
       w.Birthday, 
       w.Address, 
       w.Phone, 
       w.description, 
       w.companyID
  FROM Workers w 
  JOIN WorkerGeofences wg
    ON wg.workerID = w.id
   AND wg.isActive = 1
 WHERE w.companyID = ? 
 LIMIT ?,10