我有下面的查询,我正试图修改,以获得只有1个唯一的 PO_ID
那是最早的日期( DUE_DT
)
SELECT DISTINCT (CONVERT(CHAR(10),B.PO_DT,121)) AS 'ENTERED DATE', B.BUSINESS_UNIT, A.DEPTID, B.PO_ID,
B.PO_STATUS, B.VENDOR_ID, B.BUYER_ID,
D.RECV_REQ, (CONVERT(CHAR(10),E.DUE_DT,121)) AS 'DUE_DT'
FROM ((((PS_PO_HDR B
LEFT OUTER JOIN PS_PO_LINE_DISTRIB A ON B.BUSINESS_UNIT = A.BUSINESS_UNIT AND B.PO_ID = A.PO_ID )
LEFT OUTER JOIN PS_VENDOR C ON C.VENDOR_ID = B.VENDOR_ID )
LEFT OUTER JOIN PS_PO_LINE D ON
B.BUSINESS_UNIT = D.BUSINESS_UNIT AND B.PO_ID = D.PO_ID )
LEFT OUTER JOIN PS_PO_LINE_SHIP E ON D.BUSINESS_UNIT = E.BUSINESS_UNIT AND D.PO_ID = E.PO_ID AND D.LINE_NBR = E.LINE_NBR )
WHERE ( B.PO_STATUS IN ('A','D','O')
AND D.CANCEL_STATUS IN ('A','X','O','C')
AND E.DUE_DT BETWEEN dateadd(day, datediff(day, 0, getdate()-11), 0) AND dateadd(day, datediff(day, 0, getdate()), -3)
AND B.RECV_STATUS NOT IN ('C','M','R')
AND D.RECV_REQ = 'Y')
ORDER BY 2, 3, 4
以下是我目前获得的样本日期:
ENTERED_DT BUSINESS_UNIT DEPTID PO_ID PO_STATUS VENDOR_ID BUYER_ID RECV_REQ DUE_DT
2020-07-13 11000 624 A010035568 D UNITED NOTT Y 2020-07-13
2020-07-08 11000 659 A010035426 D SCIMED NOTT Y 2020-07-15
2020-07-08 11000 659 A010035426 D SCIMED NOTT Y 2020-07-09
2020-07-09 11000 659 A010035468 D SCIMED NOTT Y 2020-07-10
在本例中,对于第3行和第4行(其中 PO_ID
是重复的)我只想返回一行,我想返回最早日期的行( 2020-07-09
)他回来了。
我试着加上 NOT EXISTS
条件进入 WHERE
但是现在似乎过滤掉了太多的结果:
....
AND NOT EXISTS (SELECT COUNT(*) , BB.DUE_DT
FROM PS_PO_LINE_SHIP BB
LEFT OUTER JOIN PS_PO_LINE DD ON DD.BUSINESS_UNIT = BB.BUSINESS_UNIT
AND DD.PO_ID = BB.PO_ID AND DD.LINE_NBR = BB.LINE_NBR
WHERE BB.PO_ID = B.PO_ID
AND BB.BUSINESS_UNIT = B.BUSINESS_UNIT
GROUP BY BB.DUE_DT
HAVING COUNT(*) > 1)
编辑:我更改了查询以获取最小到期日( MIN((CONVERT(CHAR(10),E.DUE_DT,121)))
但是在试图计算两个 PO_DT
以及 DUE_DT
我得到了一个错误的群体。
我将此列添加到select语句中: DATEDIFF(DAY, E.DUE_DT, B.PO_DT) AS DateDiff
然而,它说它是失踪的 Group By
. 如果我将它添加到组中,那么我将重新引入重复的po\u id。有什么想法吗?
1条答案
按热度按时间xa9qqrwz1#
请看一下sql server的last\u value函数
应用到你的案例中,它看起来是这样的:
LAST_VALUE(PO_ID) OVER(ORDER BY DUE_DT) OLDEST_PO_ID