我尝试过使用CTE,但一直无法弄明白。我想将多行更改为一行,其中包含多个列。任何方式去做...
当前结果
ern CostPool
111 NULL
111 5
期望的结果
ern CostPool1 CostPool2
111 NULL 5
SELECT first_name, last_name, ern, pc, title, TitleDesc, OfficeTitle, SUM(earned) earned, fypaydate, q
, [RMTS, Support, or Partial Quarter], SUM(CAST([Staff Employment Status] AS INT)) sumEmpStatus, SUM(CAST(countisMHy AS INT)) sumcountisMHy, SUM(CAST(sumisMHy AS INT)) sumisMHy1
, SUM(CAST(sumEligible AS INT)) sumEligible1
, CostPool
FROM (
SELECT
e.first_name, e.last_name
, fss.ern
, fss.pc
, fss4.title
, u.OfficeTitle
, t.TitleDesc
, SUM(fss.gross*fss.[percent]) earned
, fss.fypaydate, fss.q
, cpayfss3, cpaypdq
, CASE WHEN cpayfss3 = cpaypdq THEN '1' --full q
WHEN cpayfss3 < cpaypdq THEN '2' --part q
ELSE 'Check'
END [RMTS, Support, or Partial Quarter]
, CASE WHEN COUNT(DISTINCT rc.isMHy) = 2 THEN '9' --split
WHEN fss.pc = 'A' AND COUNT(DISTINCT rc.isMHy) = 1 AND SUM(CAST(rc.isMHy AS INT)) > 0 THEN '1' --full
WHEN fss.pc IN ('X', 'O', 'W', 'L', 'K') AND SUM(CAST(rc.isMHy AS INT)) > 0 THEN '2' --part
ELSE '0'
END [Staff Employment Status]
, COUNT(DISTINCT rc.isMHy) countisMHy
, SUM(CAST(rc.isMHy AS INT)) sumisMHy
, SUM(CAST(rc.Eligible AS INT)) sumEligible --count distinct too?
, rc.CostPool
FROM [FIN_DataMart].[dbo].[FSSpaydetl_NEW] fss
LEFT JOIN FIN_DataMart.[dbo].[FSSRC_NEW] rc
ON fss.newrc = rc.RC
AND fss.fypaydate = rc.FY
INNER JOIN
(
SELECT ern, fypaydate, q, COUNT(DISTINCT paydate) cpayfss3
FROM [FIN_DataMart].[dbo].[FSSpaydetl_NEW]
GROUP BY fypaydate, q, ern
) fss3
ON fss.ern = fss3.ern
AND fss.fypaydate = fss3.fypaydate
AND fss.q = fss3.q
LEFT JOIN
(
SELECT COUNT(DISTINCT paydate) cpaypdq, fy, q
FROM [FIN_DataMart].[dbo].[FSSPayDateQ]
GROUP BY fy, q
) pdq
ON fss.fypaydate = pdq.fy
AND fss.q = pdq.q
LEFT JOIN
[FIN_DataMart].[dbo].FSSEmployeeInfo e
ON fss.ern = e.ern
INNER JOIN
(
SELECT DISTINCT ern, title, tlvl, fy, q, ROW_NUMBER() OVER (PARTITION BY ern, fy, q ORDER BY title DESC) rn
FROM [FIN_DataMart].[dbo].[FSSpaydetl_NEW]
) fss4
ON fss.ern = fss4.ern
AND fss.fypaydate = fss4.fy
AND fss.q = fss4.q
AND fss4.rn = '1'
LEFT JOIN
(
SELECT TCode, TLevel, GenericTDesc as TitleDesc, EED,
ROW_NUMBER() OVER (PARTITION BY TCode, TLevel ORDER BY EED DESC) rn
FROM [FIN_DataMart].[dbo].FSATitle
) t
ON fss4.title = t.TCode
AND case when fss4.tlvl = '' THEN '01' WHEN fss4.tlvl IS NULL THEN '01' ELSE fss4.tlvl END = t.TLevel
AND t.rn = 1
LEFT JOIN FIN_DataMart.[dbo].[FSSUserERN] u
ON fss.ern = u.ern
WHERE fss.fypaydate = 2015
AND fss.q = 1
AND fss.obj <> '049' AND fss.obj NOT LIKE 'x%'
AND fss.title NOT IN ('10209', '1020B', '10234', '10232', '20202', '21205', '30086', '34201')
GROUP BY
fss.ern
, fss.pc
, fss4.title
, fss.fypaydate, fss.q
, cpayfss3, cpaypdq
, rc.CostPool
, t.TitleDesc
, e.first_name
, e.last_name
, u.OfficeTitle
HAVING SUM(CAST(rc.Eligible AS INT)) >=1
) A
GROUP BY
first_name, last_name, ern, pc, title, fypaydate, q
, [RMTS, Support, or Partial Quarter]
, OfficeTitle
, TitleDesc
, CostPool
ORDER BY first_name, last_name, fypaydate, q
答案 0 :(得分:2)
您可以尝试使用 Pivot 运算符将行转换为列。
https://technet.microsoft.com/en-us/library/ms177410(v=sql.105).aspx
该方法可行,尽管您的示例数据和预期结果与您发布的大量查询不完全匹配,因此在您提供更多测试数据/预期输出(和表模式)之前,我无法提供更具体的示例。 p>