关于sql：使用CTE的相邻列表和递归查询，如何回填？

Adjacent list and recursive query using a CTE, how to backfill?

我有一个查询，它从父子表(自连接，相邻列表)生成扁平层次结构。问题是此查询为没有任何子级的级别生成 NULL。现在我的意图是"回填"这些级别以生成一个在级别列中不包含任何 NULL 值的表。我应该如何修改这个查询？

示例数据：

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

SET NOCOUNT ON;
USE Tempdb;

IF OBJECT_ID('dbo.Employees', 'U') IS NOT NULL DROP TABLE dbo.Employees;

CREATE TABLE dbo.Employees
(
empid INT NOT NULL PRIMARY KEY,
mgrid INT NULL REFERENCES dbo.Employees,
empname VARCHAR(25) NOT NULL,
salary MONEY NOT NULL,
CHECK (empid <> mgrid),
CHECK (empid > 0)
);

CREATE UNIQUE INDEX idx_unc_mgrid_empid ON dbo.Employees(mgrid, empid);

INSERT INTO dbo.Employees(empid, mgrid, empname, salary) VALUES
(1, NULL, 'David' , $10000.00),
(2, 1, 'Eitan' , $7000.00),
(3, 1, 'Ina' , $7500.00),
(4, 2, 'Seraph' , $5000.00),
(5, 2, 'Jiru' , $5500.00),
(6, 2, 'Steve' , $4500.00),
(7, 3, 'Aaron' , $5000.00),
(8, 5, 'Lilach' , $3500.00),
(9, 7, 'Rita' , $3000.00),
(10, 5, 'Sean' , $3000.00),
(11, 7, 'Gabriel', $3000.00),
(12, 9, 'Emilia' , $2000.00),
(13, 9, 'Michael', $2000.00),
(14, 9, 'Didi' , $1500.00);

@Andomar 提供的查询

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

; WITH Tree AS
(
SELECT empid
, mgrid
, 1 AS lv
, 1 AS level1
, NULL AS level2
, NULL AS level3
, NULL AS level4
, NULL AS level5
FROM Employees
WHERE mgrid IS NULL
UNION ALL
SELECT E.empid
, E.mgrid
, T.lv + 1
, T.level1
, CASE WHEN T.lv = 1 THEN E.empid ELSE t.level2 END
, CASE WHEN T.lv = 2 THEN E.empid ELSE t.level3 END
, CASE WHEN T.lv = 3 THEN E.empid ELSE t.level4 END
, CASE WHEN T.lv = 4 THEN E.empid ELSE t.level5 END
FROM Employees AS E
JOIN Tree T
ON E.mgrid = T.empid
)
SELECT *
FROM Tree
ORDER BY empid

这会产生

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

+-------+--------+----+--------+--------+--------+--------+--------+
| EMPID | MGRID | LV | LEVEL1 | LEVEL2 | LEVEL3 | LEVEL4 | LEVEL5 |
+-------+--------+----+--------+--------+--------+--------+--------+
| 1 | (NULL) | 1 | 1 | (NULL) | (NULL) | (NULL) | (NULL) |
| 2 | 1 | 2 | 1 | 2 | (NULL) | (NULL) | (NULL) |
| 3 | 1 | 2 | 1 | 3 | (NULL) | (NULL) | (NULL) |
| 4 | 2 | 3 | 1 | 2 | 4 | (NULL) | (NULL) |
| 5 | 2 | 3 | 1 | 2 | 5 | (NULL) | (NULL) |
| 6 | 2 | 3 | 1 | 2 | 6 | (NULL) | (NULL) |
| 7 | 3 | 3 | 1 | 3 | 7 | (NULL) | (NULL) |
| 8 | 5 | 4 | 1 | 2 | 5 | 8 | (NULL) |
| 9 | 7 | 4 | 1 | 3 | 7 | 9 | (NULL) |
| 10 | 5 | 4 | 1 | 2 | 5 | 10 | (NULL) |
| 11 | 7 | 4 | 1 | 3 | 7 | 11 | (NULL) |
| 12 | 9 | 5 | 1 | 3 | 7 | 9 | 12 |
| 13 | 9 | 5 | 1 | 3 | 7 | 9 | 13 |
| 14 | 9 | 5 | 1 | 3 | 7 | 9 | 14 |
+-------+--------+----+--------+--------+--------+--------+--------+

但我们的想法是实现这一目标

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

+-------+--------+----+--------+--------+--------+--------+--------+
| EMPID | MGRID | LV | LEVEL1 | LEVEL2 | LEVEL3 | LEVEL4 | LEVEL5 |
+-------+--------+----+--------+--------+--------+--------+--------+
| 1 | (NULL) | 1 | 1 | 1 | 1 | 1 | 1 |
| 2 | 1 | 2 | 1 | 2 | 2 | 2 | 2 |
| 3 | 1 | 2 | 1 | 3 | 3 | 3 | 3 |
| 4 | 2 | 3 | 1 | 2 | 4 | 4 | 4 |
| 5 | 2 | 3 | 1 | 2 | 5 | 5 | 5 |
| 6 | 2 | 3 | 1 | 2 | 6 | 6 | 6 |
| 7 | 3 | 3 | 1 | 3 | 7 | 7 | 7 |
| 8 | 5 | 4 | 1 | 2 | 5 | 8 | 8 |
| 9 | 7 | 4 | 1 | 3 | 7 | 9 | 9 |
| 10 | 5 | 4 | 1 | 2 | 5 | 10 | 10 |
| 11 | 7 | 4 | 1 | 3 | 7 | 11 | 11 |
| 12 | 9 | 5 | 1 | 3 | 7 | 9 | 12 |
| 13 | 9 | 5 | 1 | 3 | 7 | 9 | 13 |
| 14 | 9 | 5 | 1 | 3 | 7 | 9 | 14 |
+-------+--------+----+--------+--------+--------+--------+--------+

只需在树 CTE 之后添加以下内容

1
2
3
4
5
6
7
8
9

SELECT empid,mgrid,lv,
level1 = COALESCE(level1,Rn),
level2 = COALESCE(level2,Rn),
level3 = COALESCE(level3,Rn),
level4 = COALESCE(level4,Rn),
level5 = COALESCE(level5,Rn)
FROM
(SELECT empid,mgrid,lv,level1,level2,level3,level4,level5,ROW_NUMBER()OVER(ORDER BY empid) AS Rn
FROM Tree)x

enter