Question

我正在执行此查询，结果很好，但问题是它花了太长时间，有人可以告诉我如何使这个效率高（SQL显示这个查询需要2.8秒，但实际上它需要超过10秒）一开始我在3张桌子上使用连接，但是查询花费的时间比这个更长。提前致谢。

SELECT 
    ee_expert.expert_id , 
    AVG( ee_expert_rating.rating_stars ) AS total_rating, 
    ee_expert.expert_id, 
    COUNT( DISTINCT ee_expert_rating.rating_id ) AS rating_count 
FROM 
    ee_expert_rating 
        RIGHT JOIN ee_expert 
            ON ee_expert.expert_id = ee_expert_rating.expert_id 
WHERE 
    expert_verified_email =2 
    AND expert_brief_description != '' 
    AND expert_account_status =1 
    AND ee_expert.expert_id IN 
        (
            SELECT 
                expert_id 
            FROM 
                ee_expert_categories 
            WHERE 
                ee_expert_categories.category_id =5 
            GROUP BY 
                expert_id 
        ) 
GROUP BY 
    ee_expert.expert_id 
ORDER BY 
    rating_count DESC

Answer 1

这应该快一点:(删除了内联组，在这种情况下使用exists可以提供帮助。）

SELECT
    ee_expert.expert_id ,
    AVG( ee_expert_rating.rating_stars ) AS total_rating,
    COUNT( DISTINCT ee_expert_rating.rating_id ) AS rating_count
FROM
    ee_expert_rating RIGHT JOIN
    ee_expert ON ee_expert.expert_id = ee_expert_rating.expert_id
WHERE
    expert_verified_email =2 AND          
    expert_brief_description != '' AND    
    expert_account_status =1 AND          
    exists(                               
        SELECT                            
            expert_id                     
        FROM                              
            ee_expert_categories
        WHERE
            ee_expert_categories.category_id =5 and
            ee_expert_categories.expert_id=ee_expert.expert_id
    )
GROUP BY
    ee_expert.expert_id
ORDER BY
    rating_count DESC

（尝试保持IN也/没有内部组。）

Answer 2

有两件事让整个事情变得缓慢：

需要10秒，因为通过网络传输数据，并以HTML格式显示（phpMyAdmin？），它比实际语句慢。这只是猜测，但在98％的情况下都是如此。
在IN()中使用subselect总是很慢（我想因为它是针对每个父数据集递归运行的）。一个选项是使用EXISTS()

sql查询需要时间，因为我正在调查三个表

2 个答案: