Question

我有一个带有3个内部选择的SQL选择：

select t1.a, t2.b, t3.c
from (select aa as a from ...) as t1 <-- query 1
     (select bb as b from ...) as t2 <-- query 2
     (select cc as c from ...) as t3 <-- query 3
where
t1.id = t2.id and t2.id = t3.id

查询1的结果集有5行。查询2的结果集有5行。查询3的结果集有2行。

所以我的sql只显示两行，因为查询3有2行，我有一个where codition：t1.id = t2.id和t2.id = t3.id。

但最后我想看到5行，其中查询1或/和2没有任何我想得到的结果或默认值。

如果我删除where条件，那么我会得到许多记录，但这也不正确。

你能告诉我怎样才能解决它？

修改

情况就是这样：

+--------+
| client |
+--------+
     |
+------------------+
| client_benchmark |
|                  |
| + fk_client_id   |
| + fk_kpi_id      |
| + value          |
| + last_update    |
+------------------+
    |
+------+
| kpi  |
+------+

insert into kpi (ID, NAME, ALIAS_NAME...) values (1, 'kpi 1'...);
insert into kpi (ID, NAME, ALIAS_NAME...) values (2, 'kpi 2'...);
insert into kpi (ID, NAME, ALIAS_NAME...) values (3, 'kpi 3'...);
insert into KPI (ID, NAME, ALIAS_NAME...) values (4, 'kpi 4'...);
...


insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (1, 1, 1, a1_1...);
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (2, 1, 2, b1_2...);
...
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (3, 2, 1, a2_1...);
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (4, 2, 1, b2_2...);
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (5, 2, 1, c2_3...);
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (6, 2, 1, d2_4...);
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (7, 2, 1, e2_5...);
...
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (8, 3, 1, a3_1...);
insert into client_benchmark (ID, CLIENT_ID, KPI_ID, VALUE...) values (9, 3, 2, b3_2...);
...

我想要显示：

client_id | kpi_1.name | kpi_1.value | kpi_2.name | kpi_2.value | kpi_3.name | kpi_3.value
----------+------------+-------------+------------+-------------+------------+------------
1         | 'kpi 1'    | a1_1        | 'kpi 2'    | b1_2        | <null>     | <null>
2         | 'kpi 1'    | a2_1        | 'kpi 2'    | b2_2        | 'kpi 3'    | c2_3 
3         | 'kpi 1'    | a3_1        | 'kpi 2'    | b3_2        | <null>     | <null>

我需要一个快速有效的sql，尽可能少的sql join命令，我想为此创建一个视图表。

我使用postgreSQL服务器。

THX。

Answer 1

您需要LEFT JOIN：

select
    t1.a
   ,t2.b
   ,t3.c
from table1      as t1
left join table2 as t2
  on t1.id = t2.id
left join table3 as t3
  on t2.id = t3.id

除非您的子查询中有复杂的逻辑，否则如果您只是在FROM中指定表名，它可能会更具可读性。

Answer 2

JOIN？

select aa.a, bb.b, cc.c
from aa
  inner join bb on aa.id = bb.id
  inner join cc on cc.id = bb.id
where
  t1.id = t2.id and t2.id = t3.id

SQL +内部选择具有不同的行集

2 个答案: