SQL Server 2012比较记录并根据动态位置找到丢失的记录

Question

我有以下问题，我知道如何从同一个表中找到缺少的记录，但我无法知道记录从哪里丢失：

这里是主表

Location  | Role | Subrole 
A            | R1    | SR1 
A            | R1    | SR2 
A            | R1    | SR3 
B            | R1    | SR1 
B            | R1    | SR2 
B            | R1    | SR3 
C            | R1    | SR1 
C            | R1    | SR2 
D            | R1    | SR1 

位置的是主地点和所有其他位置应该是比较A.我的最终目标是拥有这样的：

MasterLocation | MasterRole | MasterSubrole | Location | Role | Subrole 
     A        | R1        | SR1          |    B     | R1   | SR1 
     A        | R1        | SR2          |    B     | R1   | SR2 
     A        | R1        | SR3          |    B     | R1   | SR3 
     A        | R1        | SR1          |    C     | R1   | SR1 
     A        | R1        | SR2          |    C     | R1   | SR2 
     A        | R1        | SR3          |    C     | R1  | MISSING OR NULL 
     A        | R1        | SR1          |    D     | R1   | SR1 
     A        | R1        | SR2          |    D     | R1   | MISSING OR NULL 
     A        | R1        | SR3          |    D     | R1   | MISSING OR NULL 

我已创建2个临时表

CREATE TABLE #LocA 
( 
    Location Varchar(1), 
    Role Varchar(2), 
    SubRole VARCHAR(20) 
) 

CREATE TABLE #AllOthers 
( 
    Location VARCHAR(1), 
    Role VARCHAR(2), 
    SubRole VARCHAR(20) 
) 

INSERT INTO #LocA 
    SELECT 
     Location, Role, SubRole 
    FROM 
     TABLE 
    WHERE 
     Location = 'A' 

INSERT INTO #AllOthers 
    SELECT 
     Location, Role, SubRole 
    FROM 
     TABLE 
    WHERE 
     Location != 'A' 

SELECT 
    A.Location AS MasterLocation, 
    A.Role AS MasterRole, 
    A.SubRole AS MasterSubrole 
    L.Location, 
    L.Role, 
    L.Subrole 
FROM 
    #LocA AS A 
LEFT JOIN 
    #Allothers AS L ON A.Role = L.Role 
        AND A.SubRole = L.Subrole 

我越来越

MasterLocation  | MasterRole  | MasterSubrole | Location | Role|Subrole 

A                       | R1              | SR1                 |     B        | R1      | SR1 
A                       | R1              | SR2                 |     B        | R1      | SR2 
A                       | R1              | SR3                 |     B        | R1      | SR3 
A                       | R1              | SR1                 |     C        | R1      | SR1 
A                       | R1              | SR2                 |     C        | R1      | SR2 
A                       | R1              | SR3                 |     NULL   | NULL   | NULL 
A                       | R1              | SR1                 |     D        | R1      | SR1 
A                       | R1              | SR2                 |     NULL   | NULL   | NULL 
A                       | R1              | SR3                 |     NULL   | NULL   | NULL 

所以我其实不知道在哪里子角色是从，C或D缺失（此表有数百个地点。）

Answer 1

交叉连接的常见用法是在这种类型的问题中，您需要显示所有可能的组合。查询的前半部分创建了该组值，然后左连接通过跨多维的复合连接来附加数据值。

select 
    m.Location as MasterLocation, 
    m.Role as MasterRole, 
    m.Subrole as MasterSubrole, 
    l.Location, 
    coalesce(t.Role, 'Missing role') as Role, 
    coalesce(t.Subrole, 'Missing subrole') as Subrole 
from 
    T as m 
    cross join 
    (select distinct Location from T where Location <> 'A') as l 
    left outer join T as t 
     on  l.Location = t.Location 
      and m.Role = t.Role 
      and m.Subrole = t.SubRole 
where 
    m.Location = 'A'; 

您可能希望避免提字面值不止一次A更在查询所以这里是一个另类。

select 
    m.Location as MasterLocation, 
    m.Role as MasterRole, 
    m.Subrole as MasterSubrole, 
    l.Location, 
    coalesce(t.Role, 'Missing role') as Role, 
    coalesce(t.Subrole, 'Missing subrole') as Subrole 
from 
    (select * from T where Location = 'A') as m 
    cross apply 
    (select distinct Location from T where Location <> m.Location) as l 
    left outer join T as t 
     on  l.Location = t.Location 
      and m.Role = t.Role 
      and m.Subrole = t.SubRole; 

如果你在数据空，你想匹配起来你可能想加入这样：

 on  l.Location = t.Location 
      and coalesce(m.Role, '!@') = coalesce(t.Role, '!@') 
      and coalesce(m.Subrole, '!@#') = coalesce(t.SubRole, '!@#'); 

http://rextester.com/MNSU54881