Question

我一直在尝试优化下面的MySQL查询无济于事。目前执行时间最长为3秒：

select count( DISTINCT v.id) from 
( 
    select sub_jobs.id from sub_jobs, main_jobs where sub_jobs.title LIKE '%abc%' and main_jobs.id=sub_jobs.parent_id 

    UNION ALL 

    select sub_jobs.id from sub_jobs, main_jobs where job_title LIKE '%abc%' and main_jobs.id=sub_jobs.parent_id 

    UNION ALL 

    select sub_jobs.id from sub_jobs, main_jobs, companies where companies.company_name LIKE '%abc%' and main_jobs.company_id=companies.id and main_jobs.id=sub_jobs.parent_id 
) 
as v

我想做什么：

在3个表中搜索三列中的关键字并返回非重复计数

基本表格结构：

1。 sub_jobs

ID
title [我想搜索此专栏]
PARENT_ID

2。 main_jobs

ID
job_title [我想搜索此专栏]
company_id [某些行没有此列的条目]

3。公司

ID
company_name [我想搜索此专栏]

EXPLAIN

------column titles------

id
select_type:
table
type
possible_keys
key
key_len
ref
rows
Extra

--- row 1 ----

id:  1
select_type: PRIMARY
table: <derived2>
ALL
NULL
NULL
NULL
NULL
102642
NULL


---- row 2 ----

id: 2
select_type: DERIVED
table: main_jobs
index
PRIMARY,id_index,id_industry_featured_index
id_index
4
NULL
34214
Using index


---- row 3 ----

id: 2
select_type: DERIVED
table: sub_jobs
ref
parent_id,parent_id_category_index
parent_id
4
myjobmag_myjobdb.main_jobs.id
1
Using where


---- row 4 ----

id: 3
select_type: UNION
table: main_jobs
ALL
PRIMARY,id_index,id_industry_featured_index
NULL
NULL
NULL
34214
Using where

---- row 5 ----

id: 3
select_type: UNION
table: sub_jobs
ref
parent_id,parent_id_category_index
parent_id
4
main_jobs.id
1
Using index

---- row 6 ----


id: 4
select_type: UNION
table: main_jobs
ALL
PRIMARY,id_index,id_industry_featured_index
NULL
NULL
NULL
34214
NULL

---- row 7 ----

id: 4
select_type: UNION
table: companies
eq_ref
PRIMARY
PRIMARY
4
main_jobs.company_id
1
Using where


---- row 8 ----

id: 4
select_type: UNION
table: sub_jobs
ref
parent_id,parent_id_category_index
parent_id
4
main_jobs.id
1
Using index


---- row 9 ----

id: NULL
select_type: UNION RESULT
table: <union2,3,4>
ALL
NULL
NULL
NULL
NULL
NULL
Using temporary

我尝试在没有DISTINCT的情况下使用UNION而没有显着的收益。

如何减少此查询的运行时间。

Answer 1

试图减少一个条件。请尝试让我知道

select count( DISTINCT v.id) from 
( 
    select sub_jobs.id 
    from sub_jobs inner join main_jobs
    on main_jobs.id=sub_jobs.parent_id  
    where (sub_jobs.title LIKE '%abc%' OR job_title LIKE '%abc%') 

    UNION ALL 

    select sub_jobs.id 
    from sub_jobs, main_jobs, companies
    where companies.company_name LIKE '%abc%' and
          main_jobs.company_id=companies.id and 
          main_jobs.id=sub_jobs.parent_id 
) 
as v

在问题更新后修改

可以组合2个这样的查询：

select count( DISTINCT v.id) from 
( 
    select sub_jobs.id 
    from sub_jobs 
         inner join main_jobs
          on main_jobs.id=sub_jobs.parent_id 
         inner join companies 
          on main_jobs.company_id=companies.id
    where companies.company_name LIKE '%abc%'
          OR
          sub_jobs.title LIKE '%abc%' 
          OR 
          job_title LIKE '%abc%'       
) 
as v

让我知道它是否有效或抛出错误

Answer 2

您可以尝试使用带有union

的3查询或代替3查询

但你应该看一下解释计划

    select count(dictinnct sub_jobs.id )
    from sub_jobs
    INNER JOIN main_jobs on main_jobs.id=sub_jobs.parent_id 
    INNER JOIN companies on main_jobs.company_id=companies.id
    where companies.company_name LIKE '%abc%' 
    OR main_jobs.job_title LIKE '%abc%'
    OR sub_jobs.title LIKE '%abc%'

UNION或UNION所有MYSQL优化

2 个答案: