Question

我们如何实现这一目标？

实际表格：

.-------.---------.-------.------.---------.
| EmpId | Project | Title | Role |  Values |
|-------|---------|-------|----- |---------|
| 1     | aaa     |xxx    |     A|      100|
| 1     | aaa     |yyy    |     B|      120|
| 1     | aaa     |zzz    |     C|       90|
.-------.---------.-------.------.---------.

目标1：

.-------.---------.-------.----.----.----.
| EmpId | Project | Title | A  | B  | C  |
|-------|---------|-------|--- |----|----|
| 1     | aaa     |xxx    | 100|null|null|
| 1     | aaa     |yyy    |null| 120|null|
| 1     | aaa     |zzz    |null|null|  90|
.-------.---------.-------.----.----.----.

目标2：

.-------.---------.----.----.----.
| EmpId | Project | A  | B  | C  |
|-------|---------|--- |----|----|
| 1     | aaa     | 100| 120|  90|
.-------.---------.----.----.----.

条件：

在目标1中，动态生成了A / B / C列。（数据透视表，不断更改的列名）。 A / B / C列实际上不是A / B / C。它是数据透视表或存储过程的结果。可以是A / B / C / D或M / N或X / Y / Z。
列Title在目标2中根本不重要。

Answer 1

只需使用聚合：

select EmpId, Project, max(A) as a, max(B) as b, max(C) as c
from t
group by EmpId, Project;

Answer 2

使用聚合。 MAX()会忽略NULL个值：

SELECT empid, project, MAX(A) as A, MAX(B) as B, MAX(C) as C
FROM mytable
GROUP BY empid, project

Demo on DB Fiddle ：

| empid | project | A   | B   | C   |
| ----- | ------- | --- | --- | --- |
| 1     | aaa     | 100 | 120 | 90  |

Answer 3

SELECT
    T1.EmpId,
    T1.Project,
    T2.A,
    T3.B,
    T4.C
FROM
    Table T1
LEFT JOIN
    Table T2 ON
        T2.EmpId=T1.EmpId
        AND T2.Project=T1.Project
        AND T2.A IS NOT NULL
LEFT JOIN
    Table T3 ON
        T3.EmpId=T1.EmpId
        AND T3.Project=T1.Project
        AND T3.B IS NOT NULL
LEFT JOIN
    Table T4 ON
        T4.EmpId=T1.EmpId
        AND T4.Project=T1.Project
        AND T4.B IS NOT NULL

Answer 4

with cte (id,pro,title,rol,val) as (
select 1,'aaa','xxx','A',100 union all
select 1,'aaa','yyy','B',120 union all
select 1,'aaa','zzz','C',90)
select id,pro,title,[a],[b],[c] from (
select * from cte ) a
pivot 
(max(val) for rol in ([a],[b],[c])) aa


with cte (id,pro,title,rol,val) as (
select 1,'aaa','xxx','A',100 union all
select 1,'aaa','yyy','B',120 union all
select 1,'aaa','zzz','C',90)
select id,pro,max([a]) A,max([b]) B,max([c]) C from (
select * from cte ) a
pivot 
(max(val) for rol in ([a],[b],[c])) aa
group by id,pro

SQL-GROUP BY-动态列

4 个答案: