表用户：
UID | NAME
------------
100 | RAVI
101 | Raju
102 | Manju
103 | Arun
表会话：
SID | DATE | TIME | STrain_ID
----------------------------------
1 | 1/1/2015 | 11:30 | 1
2 | 12/1/2015 | 12:30 | 2
3 | 25/1/2015 | 01:30 | 3
4 | 25/1/2015 | 02:30 | 4
表TRAINERS：
Train_ID | Primary | secondarry | other | primID | secID | otherID
--------------------------------------------------------------------------
1 | RAVI | | Manju | 100 | | 102
2 | | RAVI | | | 100 |
3 | | | RAVI | | | 100
4 | | Manju | Ravi | | 102 | 100
预期结果
SID | DATE | TIME | Train_ID | NAME | ROLE
-------------------------------------------------
1 | 1/1/2015 | 11:30 | 1 | RAVI |primary
1 | 1/1/2015 | 11:30 | 1 | Manju|other
2 | 12/1/2015 | 12:30 | 2 | RAVI |secondary
3 | 25/1/2015 | 01:30 | 3 | RAVI |other
4 | 25/1/2015 | 02:30 | 4 | Manju|secondary
答案 0 :(得分：1)
检查以下代码。我看到了预期结果的模式，还有一行4-Train_ID，Ravi为＆＃34;其他＆＃34;角色。正如@JamesZ已经提到过你的设计，你也必须继续努力。
SELECT S.SID, S.DATE, S.TIME, S.STrain_ID AS Train_ID ,U.NAME, CASE WHEN T.[Primary] = U.NAME THEN 'Primary'
WHEN T.[Secondary] = U.NAME THEN 'Secondary'
WHEN T.[Others] = U.NAME THEN 'Others' END AS [ROLE]
FROM #SESSION S
INNER JOIN #TRAINERS T ON S.STrain_ID = T.Train_ID
LEFT JOIN #USERS U ON (U.Name = T.[Primary])OR (U.Name = T.[Secondary]) OR (U.Name = T.Others)
如果有帮助，请将其标记为已解答。
答案 1 :(得分：0)
尝试以下查询，
select s.*,u.Name,t.src from [SESSION] s join
(select Train_ID,primID as usr,'primary' src from TRAINERS where primID is not null
union all
select Train_ID,secID, 'secondary' src from TRAINERS where secID is not null
union all
select Train_ID,otherID, 'others' src from TRAINERS where otherID is not null) t
on s.[sid]=t.train_id join USERS u on t.usr=u.[UID]