优化执行时间太长的sql查询

Question

以下sql查询要花很多时间才能返回结果

SELECT count(*)
FROM List_declarations
WHERE  Id_antenne IN (
   SELECT id
   FROM Service
   WHERE id_direction IN (
      SELECT id_dr
      FROM affectation_service_dr
      WHERE id_service = 15
   )
)

我该如何做得更好？

Answer 1

我会考虑将IN's更改为JOIN's ：

select count(*) 
from list_declarations d join service s on d.id_antenne = s.id
                         join affectation_service_dr sdr on sdr.id_dr = s.id_direction
where sdr.id_service = 15

然后，我将再次检查性能，看看在where子句和外键上使用的列上是否有任何索引。

Answer 2

尝试这个：

SELECT count(*)
FROM 
  List_declarations d
  INNER JOIN Service 
  ON d.Id_antenne = s.id

  INNER JOIN affectation_service_dr a
  s.id_direction = a.id_dr

WHERE 
  a.id_service = 15

• 它可能会有所作为，但可能不会（但通常这是编写这种性质的查询的更好方法）

不可估量的是，确保对以下内容进行索引：

affectation_service_dr.id_service 
Service.id_direction 
List_declarations.Id_antenne  

如果仍然存在无法用大量行轻松解释的性能问题（例如，占用38s，但返回的计数为20348957389956845），那么我建议您修改问题以包括计划说明（ https：// dev。 mysql.com/doc/workbench/en/wb-performance-explain.html ）

Answer 3

如果这里不关心读取未提交的数据。如果有关系，您仍然可以尝试with (nolock)删除

SELECT count(*)
FROM List_declarations LD WITH (NOLOCK)
WHERE  exists (
   SELECT id
   FROM [Service] S WITH (NOLOCK)
   inner join affectation_service_dr dr WITH (NOLOCK)
   on s.id_direction=dr.id_dr
    WHERE ld.Id_antenne=s.id and dr.id_service = 15
   )

Answer 4

我尝试了每个人的提议，但似乎没有解决方案，请求所花费的时间仍然是37s

考虑到请求，我意识到in子句中的查询是针对表list_declarations中包含大量数据的每一行执行的。 和下面的SQL代码工作镍（1S）：

declare  @table table( idantenne nvarchar(20))
insert into @table select ID from Service where id_direction in (select id_dr from affectation_service_dr where id_service = 15)
Declare @string AS Nvarchar(MAX) -- It must not be MAX if you have few numbers 
SELECT  @string = COALESCE(@string + ',', '') + idantenne
FROM   @table where idantenne IS NOT NULL
declare @req nvarchar(max)
set @req = 'SELECT * FROM List_declarations Where 1=1  and Id_antenne in (' + @string +' )'
exec (@req)