数据库服务器类型 = SQL Server 2012 SP2
这是源表。
CREATE TABLE a
(
e_id INT,
r_id INT,
d_id INT,
s_id INT
);
CREATE TABLE r
(
id INT,
NAME VARCHAR(50)
);
CREATE TABLE d
(
id INT,
calldate DATE
);
CREATE TABLE e
(
id INT,
NAME VARCHAR(50),
initial VARCHAR(5)
);
CREATE TABLE s
(
id INT,
specialty VARCHAR(50)
);
INSERT INTO a
VALUES (8,
13,
9,
13),
(8,
13,
10,
13),
(8,
13,
11,
13),
(8,
13,
12,
13),
(8,
13,
13,
13),
(8,
13,
14,
13),
(10,
13,
15,
13),
(10,
13,
16,
13),
(10,
13,
17,
13),
(16,
13,
18,
13),
(16,
13,
19,
13),
(16,
13,
20,
13),
(9,
13,
21,
13),
(9,
13,
22,
13),
(9,
13,
23,
13),
(9,
13,
24,
13),
(16,
13,
25,
13),
(16,
13,
26,
13);
INSERT INTO a
VALUES
(10,22,9,13),
(10,22,10,13),
(10,22,11,13),
(10,22,12,13),
(10,22,13,13),
(16,22,14,13),
(16,22,15,13),
(16,22,16,13),
(16,22,17,13),
(9,22,18,13),
(9,22,19,13),
(9,22,20,13),
(18,22,21,13),
(18,22,22,13),
(10,22,23,13),
(10,22,24,13),
(9,22,25,13),
(9,22,26,13);
INSERT INTO r
VALUES (13,
'Primary'),
(22,
'Secondary'),
(33,
'Prim Trd'),
(44,
'Sec Td');
INSERT INTO e
VALUES (8,
'Jack',
'JM'),
(9,
'Ace',
'AQ'),
(10,
'Mike',
'MM'),
(16,
'George',
'GH'),
(18 ,'Bill', 'BK' );
INSERT INTO s
VALUES (13,
'Pediatric'),
(28,
'EMT'),
(55,
'ER');
INSERT INTO d
VALUES (9 , '2015-01-09'),
(10 , '2015-01-10'),
(11 , '2015-01-11'),
(12 , '2015-01-12'),
(13 , '2015-01-13'),
(14 , '2015-01-14'),
(15 , '2015-01-15'),
(16 , '2015-01-16'),
(17 , '2015-01-17'),
(18 , '2015-01-18'),
(19 , '2015-01-19'),
(20 , '2015-01-20'),
(21 , '2015-01-21'),
(22 , '2015-01-22'),
(23 , '2015-01-23'),
(24 , '2015-01-24'),
(25 , '2015-01-25'),
(26 , '2015-01-26'),
(27 , '2015-01-27'),
(28 , '2015-01-28'),
(29 , '2015-01-29'),
(30 , '2015-01-30')
表 A 是链接表 D、E、R 和 S 的链接表。
表A
║ e_id │ r_id │ d_id │ s_id ║
╠══════╪══════╪══════╪══════╣
║ 8 │ 13 │ 9 │ 13 ║
║ 8 │ 13 │ 10 │ 13 ║
║ 8 │ 13 │ 11 │ 13 ║
║ 8 │ 13 │ 12 │ 13 ║
║ 8 │ 13 │ 13 │ 13 ║
║ 8 │ 13 │ 14 │ 13 ║
║ 10 │ 13 │ 15 │ 13 ║
║ 10 │ 13 │ 16 │ 13 ║
║ 10 │ 13 │ 17 │ 13 ║
║ 16 │ 13 │ 18 │ 13 ║
║ 16 │ 13 │ 19 │ 13 ║
║ 16 │ 13 │ 20 │ 13 ║
║ 9 │ 13 │ 21 │ 13 ║
║ 9 │ 13 │ 22 │ 13 ║
║ 9 │ 13 │ 23 │ 13 ║
║ 9 │ 13 │ 24 │ 13 ║
║ 16 │ 13 │ 25 │ 13 ║
║ 16 │ 13 │ 26 │ 13 ║
║ 10 │ 22 │ 9 │ 13 ║
║ 10 │ 22 │ 10 │ 13 ║
║ 10 │ 22 │ 11 │ 13 ║
║ 10 │ 22 │ 12 │ 13 ║
║ 10 │ 22 │ 13 │ 13 ║
║ 16 │ 22 │ 14 │ 13 ║
║ 16 │ 22 │ 15 │ 13 ║
║ 16 │ 22 │ 16 │ 13 ║
║ 16 │ 22 │ 17 │ 13 ║
║ 9 │ 22 │ 18 │ 13 ║
║ 9 │ 22 │ 19 │ 13 ║
║ 9 │ 22 │ 20 │ 13 ║
║ 18 │ 22 │ 21 │ 13 ║
║ 18 │ 22 │ 22 │ 13 ║
║ 10 │ 22 │ 23 │ 13 ║
║ 10 │ 22 │ 24 │ 13 ║
║ 9 │ 22 │ 25 │ 13 ║
║ 9 │ 22 │ 26 │ 13 ║
表D
║ id │ calldate ║
╠════╪════════════╣
║ 9 │ 2015-01-09 ║
║ 10 │ 2015-01-10 ║
║ 11 │ 2015-01-11 ║
║ 12 │ 2015-01-12 ║
║ 13 │ 2015-01-13 ║
║ 14 │ 2015-01-14 ║
║ 15 │ 2015-01-15 ║
║ 16 │ 2015-01-16 ║
║ 17 │ 2015-01-17 ║
║ 18 │ 2015-01-18 ║
║ 19 │ 2015-01-19 ║
║ 20 │ 2015-01-20 ║
║ 21 │ 2015-01-21 ║
║ 22 │ 2015-01-22 ║
║ 23 │ 2015-01-23 ║
║ 24 │ 2015-01-24 ║
║ 25 │ 2015-01-25 ║
║ 26 │ 2015-01-26 ║
║ 27 │ 2015-01-27 ║
║ 28 │ 2015-01-28 ║
║ 29 │ 2015-01-29 ║
║ 30 │ 2015-01-30 ║
表E
╔════╤════════╤═════════╗
║ id │ name │ initial ║
╠════╪════════╪═════════╣
║ 8 │ Jack │ JM ║
║ 9 │ Ace │ AQ ║
║ 10 │ Mike │ MM ║
║ 16 │ George │ GH ║
║ 18 │ Bill │ BK ║
╚════╧════════╧═════════╝
表R
╔════╤═══════════╗
║ id │ name ║
╠════╪═══════════╣
║ 13 │ Primary ║
║ 22 │ Secondary ║
║ 33 │ Prim Trd ║
║ 44 │ Sec Td ║
╚════╧═══════════╝
表S
|ID | Specialty |
|---+-----------|
|13 | Pediatric |
|28 | EMT |
|55 | ER |
这是我认为可行但会产生错误的查询:
SELECT d.calldate,
Max(CASE
WHEN s.id = 13
AND r.NAME = 'Primary' THEN ( e.NAME )
END) AS [Primary],
Max(CASE
WHEN s.id = 13
AND r.NAME = 'Secondary' THEN ( e.NAME )
END) AS [Secondary],
Max(CASE
WHEN s.id = 13
AND r.NAME = 'Primary'
AND ( ( e.id = (SELECT e1.id
FROM a a1
INNER JOIN e e1
ON e1.id = a1.e_id
INNER JOIN d d1
ON d1.id = a1.d_id
INNER JOIN s s1
ON s1.id = a1.s_id
INNER JOIN r r1
ON r1.id = a1.r_id
WHERE d1.calldate = CONVERT(DATE,
Dateadd(dd, -1,
d.calldate),
101)
AND r1.NAME = 'Primary'
AND s1.id = 13) )
AND ( e.id = (SELECT e1.id
FROM a a1
INNER JOIN e e1
ON e1.id = a1.e_id
INNER JOIN d d1
ON d1.id = a1.d_id
INNER JOIN s s1
ON s1.id = a1.s_id
INNER JOIN r r1
ON r1.id = a1.r_id
WHERE d1.calldate = CONVERT(DATE, Dateadd
(dd, 1,
d.calldate),
101
)
AND r1.NAME = 'Primary'
AND s1.id = 13) ) ) THEN ' '
ELSE e.initial
END) AS [PrimaryINIT],
Max(CASE
WHEN s.id = 13
AND r.NAME = 'Secondary'
AND ( e.id = (SELECT e1.id
FROM a a1
INNER JOIN e e1
ON e1.id = a1.e_id
INNER JOIN d d1
ON d1.id = a1.d_id
INNER JOIN s s1
ON s1.id = a1.s_id
INNER JOIN r r1
ON r1.id = a1.r_id
WHERE d1.calldate = CONVERT(DATE, Dateadd(dd, -1,
d.calldate),
101)
AND r1.NAME = 'Secondary'
AND s1.id = 13)
AND ( e.id = (SELECT e1.id
FROM a a1
INNER JOIN e e1
ON e1.id = a1.e_id
INNER JOIN d d1
ON d1.id = a1.d_id
INNER JOIN s s1
ON s1.id = a1.s_id
INNER JOIN r r1
ON r1.id = a1.r_id
WHERE d1.calldate = CONVERT(DATE, Dateadd(dd, 1,
d.calldate),
101
)
AND r1.NAME = 'Secondary'
AND s1.id = 13) ) ) THEN ' '
ELSE e.initial
END) AS [SecondaryINIT]
FROM a
INNER JOIN e
ON e.id = a.e_id
INNER JOIN d
ON d.id = a.d_id
INNER JOIN s
ON s.id = a.s_id
INNER JOIN r
ON r.id = a.r_id
WHERE Datepart(mm, d.calldate) = 1
AND Datepart(year, d.calldate) = 2015
GROUP BY d.calldate
这是我们从查询中寻找的结果。
(请注意，初始列仅在数据更改前后返回列值):
查询的所需结果。
╔════════════╤═════════╤═══════════╤═════════╤═══════════╗
║ callDate │ Primary │ Secondary │ PrimINIT│ SeconINIT ║
╠════════════╪═════════╪═══════════╪═════════╪═══════════╣
║ 2015-01-09 │ Jack │ Mike │ JM │ MM ║
║ 2015-01-10 │ Jack │ Mike │ │ ║
║ 2015-01-11 │ Jack │ Mike │ │ ║
║ 2015-01-12 │ Jack │ Mike │ │ ║
║ 2015-01-13 │ Jack │ Mike │ │ MM ║
║ 2015-01-14 │ Jack │ George │ JM │ GH ║
║ 2015-01-15 │ Mike │ George │ MM │ ║
║ 2015-01-16 │ Mike │ George │ │ ║
║ 2015-01-17 │ Mike │ George │ MM │ GH ║
║ 2015-01-18 │ George │ Ace │ GH │ AQ ║
║ 2015-01-19 │ George │ Ace │ │ ║
║ 2015-01-20 │ George │ Ace │ GH │ AQ ║
║ 2015-01-21 │ Ace │ Bill │ AQ │ BK ║
║ 2015-01-22 │ Ace │ Bill │ │ BK ║
║ 2015-01-23 │ Ace │ Mike │ │ MM ║
║ 2015-01-24 │ Ace │ Mike │ AQ │ MM ║
║ 2015-01-25 │ George │ Ace │ GH │ AQ ║
║ 2015-01-26 │ George │ Ace │ GH │ AQ ║
╚════════════╧═════════╧═══════════╧═════════╧═══════════╝
最佳答案
像这样尝试:
WITH Pivoted AS
(
SELECT d.calldate AS calldate
,MAX(CASE WHEN r.Name='Primary' THEN e.NAME END) AS [Primary]
,MAX(CASE WHEN r.Name='Secondary' THEN e.NAME END) AS [Secondary]
,MAX(CASE WHEN r.Name='Primary' THEN e.initial END) AS [PrimINIT]
,MAX(CASE WHEN r.Name='Secondary' THEN e.initial END) AS [SeconINIT]
FROM a
INNER JOIN e ON a.e_id=e.id
INNER JOIN r ON a.r_id=r.id
INNER JOIN d ON a.d_id=d.id
INNER JOIN s ON a.s_id=s.id
GROUP BY d.calldate
)
,WithNeighbours AS
(
SELECT *
,ISNULL(LAG([Primary]) OVER(ORDER BY calldate,[Primary]),'') AS PrevPrim
,ISNULL(LAG([Secondary]) OVER(ORDER BY calldate,[Secondary]),'') AS PrevSecon
,ISNULL(LEAD([Primary]) OVER(ORDER BY calldate,[Primary]),'') AS NxtPrim
,ISNULL(LEAD([Secondary]) OVER(ORDER BY calldate,[Secondary]),'') AS NxtSecon
FROM Pivoted
)
SELECT calldate
,[Primary]
,[Secondary]
,CASE WHEN [Primary]<>PrevPrim OR [Primary]<>NxtPrim THEN PrimINIT ELSE '' END AS PrimINIT
,CASE WHEN [Secondary]<>PrevSecon OR [Secondary]<>NxtSecon THEN SeconINIT ELSE '' END AS SeconINIT
FROM WithNeighbours
ORDER BY calldate
给定数据的结果
+------------+---------+-----------+----------+-----------+
| calldate | Primary | Secondary | PrimINIT | SeconINIT |
+------------+---------+-----------+----------+-----------+
| 2015-01-09 | Jack | Mike | JM | MM |
+------------+---------+-----------+----------+-----------+
| 2015-01-10 | Jack | Mike | | |
+------------+---------+-----------+----------+-----------+
| 2015-01-11 | Jack | Mike | | |
+------------+---------+-----------+----------+-----------+
| 2015-01-12 | Jack | Mike | | |
+------------+---------+-----------+----------+-----------+
| 2015-01-13 | Jack | Mike | | MM |
+------------+---------+-----------+----------+-----------+
| 2015-01-14 | Jack | George | JM | GH |
+------------+---------+-----------+----------+-----------+
| 2015-01-15 | Mike | George | MM | |
+------------+---------+-----------+----------+-----------+
| 2015-01-16 | Mike | George | | |
+------------+---------+-----------+----------+-----------+
| 2015-01-17 | Mike | George | MM | GH |
+------------+---------+-----------+----------+-----------+
| 2015-01-18 | George | Ace | GH | AQ |
+------------+---------+-----------+----------+-----------+
| 2015-01-19 | George | Ace | | |
+------------+---------+-----------+----------+-----------+
| 2015-01-20 | George | Ace | GH | AQ |
+------------+---------+-----------+----------+-----------+
| 2015-01-21 | Ace | Bill | AQ | BK |
+------------+---------+-----------+----------+-----------+
| 2015-01-22 | Ace | Bill | | BK |
+------------+---------+-----------+----------+-----------+
| 2015-01-23 | Ace | Mike | | MM |
+------------+---------+-----------+----------+-----------+
| 2015-01-24 | Ace | Mike | AQ | MM |
+------------+---------+-----------+----------+-----------+
| 2015-01-25 | George | Ace | GH | AQ |
+------------+---------+-----------+----------+-----------+
| 2015-01-26 | George | Ace | GH | AQ |
+------------+---------+-----------+----------+-----------+
关于sql-server - 寻找查询以在数据更改时返回数据值，否则返回空间 ' '，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/38211905/