Select 多个表中的多个计数 (*) 具有单个查询

Question

我有 3 个 tables:

基本

id	姓名	描述
2	名字 1	描述2
3	名字2	描述3

链接A

id	linkA_ID
2	344
3	3221
2	6642
3	2312
2	323

链接B

id	linkB_ID
2	8287
3	42466
2	616422
3	531
2	2555
2	8592
3	1122
2	33345

我想获得如下 table 结果：

id	姓名	描述	linkA_count	linkB_count
2	名字 1	描述2	3	2
3	名字2	描述3	5	3

我的查询：

SELECT
     a.id
    ,a.name
    ,a.description
    ,COUNT(b.linkA_ID) AS linkA_count
    ,COUNT(c.linkB_ID) AS linkb_count
FROM
    basic a 
    JOIN linkA b on (a.id = b.id)
    JOIN linkb c on (a.id = c.id)
GROUP BY
    a.id
    ,a.name
    ,a.description

查询结果是 linkA 的计数始终与 linkB 相同

Answer 1

试试这个（使用子查询）

SELECT
 basic.id
,basic.name
,basic.description
,(select Count(linkA_ID) from LinkA  where LinkA.id=basic.id) as LinkACount
,(select Count(linkB_ID) from LinkB  where LinkB.id=basic.id) as LinkBCount FROM basic

方法 2（尝试 CTE）

   with a as(select id,Count(linkA_ID)LinkACount from LinkA group by id)
, b as (select id,Count(linkB_ID)LinkBCount from LinkB group by id)
select basic.id,a.LinkACount,b.linkBCount
from basic 
join a on (a.id=basic.id)
join b on (b.id=basic.id)

Answer 2

一种更传统的方法是使用“派生表”（子查询），以便在联接乘以行之前执行计数。使用左连接允许查询返回 basic 中的所有 id，即使在两个连接表中都没有相关行也是如此。

select
      basic.id
    , coalesce(a.LinkACount,0) LinkACount
    , coalesce(b.linkBCount,0) linkBCount
from basic 
left join (
     select id, Count(linkA_ID) LinkACount from LinkA group by id
     ) as a on a.id=basic.id
left join (
     select id, Count(linkB_ID) LinkBCount from LinkB group by id
     ) as b on b.id=basic.id

Answer 3

如果您的 table 中只有 select，您就会明白为什么您的查询无法工作。

 SELECT    
    *
 FROM 
    basic a 
    JOIN linkA b on (a.id = b.id)   
    JOIN linkb c on (a.id = c.id)    
 WHERE a.ID = 3

=> 只需在计数中使用 distinct

SELECT
     a.id
    ,a.name
    ,a.description
    ,COUNT(DISTINCT(b.linkA_ID)) AS linkA_count
    ,COUNT(DISTINCT(c.linkB_ID)) AS linkb_count
FROM
    basic a 
    JOIN linkA b on (a.id = b.id)
    JOIN linkb c on (a.id = c.id)
GROUP BY
    a.id
    ,a.name
    ,a.description

Select 多个表中的多个计数 (*) 具有单个查询

Select multiple count(*) in multiple tables with single query

sql

hive

count

hiveql