GROUP_CONCAT与独特和子查询性能问题

Question

我想在MYSQL数据库中执行此查询。

SELECT FirstName,LastName 
    CONCAT(COALESCE(Address1, ''), 
    COALESCE(Address2, '')) AS Address2, 
    (SELECT GROUP_CONCAT(EmpDept ' ') FROM EMP_DEPT 
    USE INDEX (EMP_DEPT_IDX) 
    WHERE EMP_DEPT.EMP_ID = EMP.EMP_ID) AS DEPT, 
    COUNT(PIN) AS PIN, 
    MIN(SALARY) AS SALARY, 
    GROUP_CONCAT(DISTINCT PAY_CODE SEPARATOR ' ') AS PAYROLL_CODE 
FROM EMP WHERE HOUSE='YELLOW' AND HOUSE_AREA='220' 
    AND (JOIN_YEAR = '2011' OR JOIN_YEAR = '2012') 
    AND EMP_BAND='G' 
    AND (FRESHER = 'N' OR FRESHER = 'FALSE') 
    AND PIN IS NOT NULL 
    AND SALARY > 0 
GROUP BY FirstName,LastName Address3, Address2, DEPT 

该查询工作正常，我得到所需的结果集。

只有问题，没有对这两行代码巨大的性能损失，采用GROUP_CONCAT写，

(SELECT GROUP_CONCAT(EmpDept ' ') FROM EMP_DEPT 
    USE INDEX (EMP_DEPT_IDX) 
    WHERE EMP_DEPT.EMP_ID = EMP.EMP_ID) AS DEPT 


GROUP_CONCAT(DISTINCT PAY_CODE SEPARATOR ' ') AS PAYROLL_CODE 

我尝试添加索引，但仍然没有运气。（仅这些子查询正在各地30-40sec执行）

我试图实现连接，但我是新来的mysql没有运气..

谁能帮助我在做我的查询执行得更快？

Answer 1

尝试与子查询外部联接替换相关子查询：

SELECT FirstName,LastName 
    CONCAT(COALESCE(Address1, ''), 
    COALESCE(Address2, '')) AS Address2, 
    COALESCE(DEPT, '') AS DEPT, 
    COUNT(PIN) AS PIN, 
    MIN(SALARY) AS SALARY, 
    GROUP_CONCAT(DISTINCT PAY_CODE SEPARATOR ' ') AS PAYROLL_CODE 
FROM EMP 
LEFT JOIN (SELECT EMP_ID, GROUP_CONCAT(EmpDept SEPARATOR ' ') AS DEPT 
      FROM EMP_DEPT 
      GROUP BY EMP_ID) AS EMP_DEPT ON EMP_DEPT.EMP_ID = EMP.EMP_ID 
WHERE HOUSE='YELLOW' AND HOUSE_AREA='220' 
    AND (JOIN_YEAR = '2011' OR JOIN_YEAR = '2012') 
    AND EMP_BAND='G' 
    AND (FRESHER = 'N' OR FRESHER = 'FALSE') 
    AND PIN IS NOT NULL 
    AND SALARY > 0 
GROUP BY EMP.EMP_ID 

另外，我改变了最终GROUP BY使用EMP_ID列，而不是返回的列值。

也可能只是更改最后的GROUP BY将解决您的问题。按索引列分组应该更高效，因此请确保您有EMP_ID的索引。

Answer 2

该子查询的最佳索引是EMP_DEPT(EMPT_ID, EmpDept)。那是你定义的索引吗？第二个group_concat()不能利用索引。

如果删除这两行，查询速度会快多少？