3

我有一个 ID 和职位表

CREATE TABLE #MissingSequence (ID INT NOT NULL, Position INT NOT NULL)
INSERT INTO #MissingSequence (ID,Position)
SELECT 36,1
UNION ALL SELECT 36,2
UNION ALL SELECT 36,3
UNION ALL SELECT 36,4
UNION ALL SELECT 36,5
UNION ALL SELECT 36,6
UNION ALL SELECT 44,1
UNION ALL SELECT 44,3
UNION ALL SELECT 44,4
UNION ALL SELECT 44,5
UNION ALL SELECT 44,6

我要查找的是，在这种情况下，ID 的位置序列中是否有任何中断，即 44,1 和 44,3 之间的中断

我设法一起解析：

SELECT  l.ID
    ,Start_Position = MIN(l.Position) + 1
    ,Stop_Position = MIN(fr.Position) - 1
FROM #MissingSequence l
LEFT JOIN #MissingSequence r 
    ON l.Position = r.Position - 1
LEFT JOIN #MissingSequence fr 
    ON l.Position < fr.Position
WHERE r.Position IS NULL
    AND fr.Position IS NOT NULL
GROUP BY l.ID

但如果有多个 ID 值，它就不起作用。如果只有一个 ID，44 存在，它确实有效。

想法、意见、建议？

谢谢！

4

3 回答 3

8

左自联接是一种很好的直觉，但我认为聚合不会削减它，当然您需要在自联接中包含匹配 ID 子句。

这是一个使用空左连接思想的（符合 ANSI 标准的）版本，选择顶行和底行并检查它们之间没有任何内容：

SELECT
    above.ID AS ID, below.Position+1 AS Start_Position, above.Position-1 AS End_Position
FROM MissingSequence AS above
JOIN MissingSequence AS below
    ON below.ID=above.ID AND below.Position<above.Position-1
LEFT JOIN MissingSequence AS inbetween
    ON inbetween.ID=below.ID AND inbetween.Position BETWEEN below.Position+1 AND above.Position-1
WHERE inbetween.ID IS NULL;

+----+----------------+--------------+
| ID | Start_Position | End_Position |
+----+----------------+--------------+
| 44 |              2 |            2 | 
+----+----------------+--------------+
于 2009-04-07T20:31:46.760 回答
2

此查询发现单据，希望有用；如果您使用的是 SQL 2005，则可以使用 CTE

SELECT ID, Position + 1
FROM #MissingSequence t1
WHERE (Position + 1) NOT IN (SELECT Position FROM #MissingSequence t2 WHERE t1.ID = t2.ID)
AND Position <> (SELECT MAX(Position) FROM #MissingSequence t2 WHERE t1.ID = t2.ID)
于 2009-04-07T20:19:00.113 回答
0 
create database testing
use testing;
create table sequence (
    id int not null primary key
);

insert into sequence(id) values
    (1), (2), (3), (4), (6), (7), (8), (9),
    (10), (15), (16), (17), (18), (19), (20);

select * from sequence

Create PROCEDURE test_proce(@mode varchar(50))   
AS
BEGIN
    declare @se int;
    set @se=0;
    set @se=(
        select top 1 t.id + 1 
        from sequence t 
        left join sequence x on x.id = t.id + 1 
        where x.id is null 
        order by t.id
    );
    select * from sequence where id<@se;
END

exec test_proce 'mode'
于 2014-02-20T08:43:49.060 回答