MySQL 查询需要很长时间才能完成答案

【问题标题】：MySQL query taking too long to finishMySQL 查询需要很长时间才能完成
【发布时间】：2015-06-05 02:42:41
【问题描述】：

我正在使用 MySQL 数据库和以下查询：

SELECT
    customer_id,
    customer_code, 
    customer_name,  
    customer_address,
    customer_home_town
FROM  
    customers
    INNER JOIN (
        SELECT 
            issue_entry_customer_id,
            SUM(issue_entry_copies
              +issue_entry_subscribe
              +issue_entry_freecopies) AS col_copies 
        FROM 
            issue_entry 
        WHERE 
            issue_entry_deleted_status = 0  AND
            DATE_FORMAT(issue_entry_date,'%Y-%m') 
                = '".dateDatabaseFormat($issue_date)."'     
        GROUP BY 
            issue_entry_customer_id
    ) col_table 
        ON issue_entry_customer_id = customer_id
            AND customer_code = 1
            AND col_table.col_copies != 0
    ORDER BY
        customer_id ASC

issue_entry 表有超过 500 000 行。

如何提高性能？

【问题讨论】：

只有完整的查询

标签： mysql database

【解决方案1】：

table_name alias INNER JOIN table_name2 alias ON ...

这就是 INNER JOIN 查询的方式。为什么要在 INNER JOIN 语句之后进行选择？

【讨论】：

【解决方案2】：

这是您的查询：

SELECT c.customer_id, c.customer_code, c.customer_name, 
       c.customer_address, c.customer_home_town
FROM customers c INNER JOIN
      (SELECT issue_entry_customer_id, 
              SUM(issue_entry_copies + issue_entry_subcripe + issue_entry_freecopies
              ) AS col_copies 
       FROM issue_entry ie
       WHERE issue_entry_deleted_status = 0  AND
             DATE_FORMAT(issue_entry_date, '%Y-%m')  = '".dateDatabaseFormat($issue_date)."'                 
      GROUP BY issue_entry_customer_id
     ) col_table
     ON col_table.issue_entry_customer_id = c.customer_id AND
        c.customer_code = 1 AND
        col_table.col_copies <> 0  
ORDER BY c.customer_id ASC;

我的建议是更改查询的整个格式：

SELECT c.customer_id, c.customer_code, c.customer_name, 
       c.customer_address, c.customer_home_town
FROM customers c
WHERE c.customer_code = 1 AND
      EXISTS (SELECT 1
              FROM issue_entry ie
              WHERE ie.issue_entry_customer_id = customer_id AND  
                    DATE_FORMAT(issue_entry_date, '%Y-%m')  = '".dateDatabaseFormat($issue_date)."' AND
                    ie.issue_entry_deleted_status = 0 AND
                    (issue_entry_copies > 0 OR
                     issue_entry_subcripe > 0 OR
                     issue_entry_freecopies
                    )
             )
ORDER BY customer_id ASC;

注意：这假设三列永远不会是负数。

对于此查询，您需要以下索引：customers(customer_code, customer_id) 和 issue_entry(issue_entry_customer_id, issue_entry_deleted_status)。

【讨论】：

【解决方案3】：

您的查询使用了一些别名以获得更好的可读性。

select
c.customer_id, 
c.customer_code, 
c.customer_name,  
c.customer_address,
c.customer_home_town
from customers c
inner join (
 SELECT 
 issue_entry_customer_id, 
 SUM(issue_entry_copies+issue_entry_subcripe+issue_entry_freecpoies) AS col_copies 
 FROM issue_entry 
 WHERE 
 issue_entry_deleted_status = 0  AND
 DATE_FORMAT(issue_entry_date,'%Y-%m') = '".dateDatabaseFormat($issue_date)."'                 
 GROUP BY 
 issue_entry_customer_id
)x
on x.issue_entry_customer_id = c.customer_id
and c.customer_code   = 1
and x.col_copies != 0  
order BY customer_id ASC

现在对于大数据，您当然需要添加一些索引来提升查询。

alter table issue_entry add index i_del_date(issue_entry_deleted_status,issue_entry_date);
alter table issue_entry add index i_cid_idx(issue_entry_customer_id);
alter table customers add index c_code_idx(customer_code);

如果客户表中的customer_id 不是主键，您可能还需要对其进行索引

alter table customers add index c_customer_id_idx(customer_id);

另请注意，即使 issue_entry_date 被编入索引，它也可能无法使用索引，因为您正在格式化日期。

确保在应用索引之前备份表。

【讨论】：