Question

有3张桌子

表A

id | value
-----------
   |

表B

id|value|A_id(fk to A)
--------------
  |     |

表C

id|value|B_id(FK to B)|timestamp
--------------------------------
  |     |             |

我编写了一个查询，使用以下查询找出所有最新的不同C值

select A.id, B.id, C.timestamp, C.value
from A,B,C 
where A.id = B.A_id
and B.id = C.B_id
where C.value in (select distinct value from C c2 where c2.value = c.value and c2.value is not null)
and c.timestamp = (select max(timestamp) from C c3 where c3.value = c.value);

除了ID之外，其他列都没有索引。此时此查询需要大约2小时或更长时间才能运行，因为不同C值的数量是221000条记录。有没有一种有效的方法呢？

Answer 1

将在主查询中的每一行运行查询内的子查询在主查询中包含大数据时，这将是一个性能反模式（您有2个子查询）。

您需要一个最大组，可以使用自我左连接来实现。

SELECT A.id a_id, B.id b_id, C1.timestamp, C1.value
From C C1
INNER JOIN B on B.id = C1.b_id
INNER JOIN A on A.id = B.A_id
LEFT JOIN C C2 on C1.value = C2.value 
and 
C1.timstamp < C2.timestamp
WHERE C1.value IS NOT NULL
and C2.id IS NULL

Answer 2

SELECT distinct A.id, B.id, c.timestamp, c.value FROM 
(
SELECT  c.value,  MAX(c.timestamp) AS max_timestamp FROM c
WHERE NOT c.value IS NULL
GROUP BY  c.value) c1 INNER JOIN c ON c1.value = c.value AND c1.max_timestamp = c.timestamp
    inner join b ON B.id = C.B_id
    inner join a ON A.id = B.A_id

Answer 3

尝试其中之一。我想我理解你的联系。希望能帮助到你。第二个查询可能会更快，因为它首先尝试收集最大时间戳，然后加入其他表ID

select A.id, B.id.C.timestamp,c.value
from A 
inner join B on B.B_id = A.id
inner join C on C.C_id = B._B_id
where C.value in (select distinct value from C c2 where c2.value = c.value and c2.value is not null)
and c.timestamp = (select max(timestamp) from C c3 where c3.value = c.value);

WITH

A_ID AS (
SELECT A.id, B.id
from A 
inner join B on B.A_id = A.id
)
, C_ID AS (
SELECT C_ID, value, max(timestamp) 
from C
where value in (select distinct value from C where value is not null)
)
SELECT a.id, a.b_id, c.B_id, c.timestamp, c.value
FROM A_ID a
inner join C_ID c on c.B_ID = a.B_ID
order by a.id

在以下情况下需要有效的查询

3 个答案: