查找SQL Server 自增ID值不連續記錄

在很多的時候，我們會在數據庫的表中設置一個字段：ID，這個ID是一個IDENTITY，也就是說這是一個自增ID。當並發量很大並且這個字段不是主鍵的時候，就有可能會讓這個值重復；或者在某些情況（例如插入數據的時候出錯，或者是用戶使用了Delete刪除了記錄）下會讓ID值不是連續的，比如1,2,3,5,6,7,10，那么在中間就斷了幾個數據，那么我們希望能在數據中找出這些相關的記錄，我希望找出的記錄是3,5,7,10，通過這些記錄可以查看這些記錄的規律來分析或者統計；又或者我需要知道那些ID值是沒有的：4,8,9。

解決辦法的核心思想是： 獲取到當前記錄的下一條記錄的ID值，再判斷這兩個ID值是否差值為1，如果不為1那就表示數據不連續了

執行下面的語句生成測試表和測試記錄

--生成測試數據
if exists (select * from sysobjects where id = OBJECT_ID('[t_IDNotContinuous]') and OBJECTPROPERTY(id, 'IsUserTable') = 1) 
DROP TABLE [t_IDNotContinuous]

CREATE TABLE [t_IDNotContinuous] (
[ID] [int]  IDENTITY (1, 1)  NOT NULL,
[ValuesString] [nchar]  (10) NULL)

SET IDENTITY_INSERT [t_IDNotContinuous] ON

INSERT [t_IDNotContinuous] ([ID],[ValuesString]) VALUES ( 1,'test')
INSERT [t_IDNotContinuous] ([ID],[ValuesString]) VALUES ( 2,'test')
INSERT [t_IDNotContinuous] ([ID],[ValuesString]) VALUES ( 3,'test')
INSERT [t_IDNotContinuous] ([ID],[ValuesString]) VALUES ( 5,'test')
INSERT [t_IDNotContinuous] ([ID],[ValuesString]) VALUES ( 6,'test')
INSERT [t_IDNotContinuous] ([ID],[ValuesString]) VALUES ( 7,'test')
INSERT [t_IDNotContinuous] ([ID],[ValuesString]) VALUES ( 10,'test')

SET IDENTITY_INSERT [t_IDNotContinuous] OFF

select * from [t_IDNotContinuous]

（圖1：測試表）

--拿到當前記錄的下一個記錄進行連接
select ID,new_ID
into [t_IDNotContinuous_temp]
from (
select ID,new_ID = (
select top 1 ID from [t_IDNotContinuous]
where ID=(select min(ID) from [t_IDNotContinuous] where ID>a.ID)
)
from [t_IDNotContinuous] as a
) as b

select * from [t_IDNotContinuous_temp]

(圖2：錯位記錄)

--不連續的前前后后記錄
select * 
from [t_IDNotContinuous_temp]
where ID <> new_ID - 1


--查詢原始記錄
select a.* from [t_IDNotContinuous] as a
inner join (select * 
from [t_IDNotContinuous_temp]
where ID <> new_ID - 1) as b
on a.ID >= b.ID and a.ID <=b.new_ID
order by a.ID

（圖3：效果）

 

補充1：如果這個ID字段不是主鍵，那么就會有ID值重復的情況（有可能是一些誤操作，之前就有遇到過）那么就需要top 1來處理。但是當前這種情況下可以使用下面的簡化語句

select a.id as oid, nid = 
(select min(id) from t_IDNotContinuous b where b.id > a.id) 
from t_IDNotContinuous a

補充2：缺失ID值列表，

--方法一：找出上一條記錄+1，再比較大小
select (select max(id)+1 
from [t_IDNotContinuous] 
where id<a.id) as beginId,
(id-1) as endId
from [t_IDNotContinuous] a
where
a.id>(select max(id)+1 from [t_IDNotContinuous] where id<a.id)

 

（圖4：效果）

--方法二：全部+1，再判斷在原來記錄中找不到
select beginId,
(select min(id)-1 from [t_IDNotContinuous] where id > beginId) as endId 
from (  
select id+1 as beginId from [t_IDNotContinuous]
where id+1 not in 
(select id from [t_IDNotContinuous]) 
and id < (select max(id) from [t_IDNotContinuous])  
) as t