0
我有一个很大的查询，相当大。即使我指定了“DISTINCT”，该查询也会在MATERIAL列上不断生成多个结果。我想通过指定'DISTINCT'并将我需要的值相加，以便每个材料编号产生一行。任何人都可以帮助我看到我在这里做错了什么。我意识到最后的子查询在这一点上可能是多余的 - 我正在测试，这正是我得到的结果。根据SELECT字段中的一个值在大型查询中选择最近的记录
提前致谢。
DECLARE @daynumber as int
SET @daynumber = CASE
WHEN DATENAME(dw,GETDATE()) = 'Monday'
THEN 56
WHEN DATENAME(dw,GETDATE()) = 'Tuesday'
THEN 57
WHEN DATENAME(dw,GETDATE()) = 'Wednesday'
THEN 58
WHEN DATENAME(dw,GETDATE()) = 'Thursday'
THEN 59
WHEN DATENAME(dw,GETDATE()) = 'Friday'
THEN 60
END;
SELECT DISTINCT MATERIAL,
PLANT_CODE,
STOCK_CATEGORY,
material_desc AS 'MATERIAL_DESC',
division_id AS 'DIVISION_ID',
UPPER(division_desc) AS 'DIVISION_DESC',
UPPER(gender_id) AS 'GENDER_ID',
UPPER(gender_desc) AS 'GENDER_DESC',
DISTRIBUTION_VERSION_CODE,
PERIOD_CODE,
REQUIREMENTS_DATE,
VERSION_IND_FLAG,
TECHNICAL_INDEX,
SIZE_LITERAL,
[ORIGINAL FCST QTY],
WITHDRAWN_QUANTITY,
[REM PLAN QTY],
FUTURE_FCST,
TOTAL_OH_INV
FROM (
SELECT P1.PLANT_CODE,
P1.STOCK_CATEGORY,
P1.MATERIAL,
M1.material_desc AS 'MATERIAL_DESC',
M1.division_id AS 'DIVISION_ID',
UPPER(M1.division_desc) AS 'DIVISION_DESC',
UPPER(M1.gender_id) AS 'GENDER_ID',
UPPER(M1.gender_desc) AS 'GENDER_DESC',
P1.DISTRIBUTION_VERSION_CODE,
P1.PERIOD_CODE,
P1.REQUIREMENTS_DATE,
P1.VERSION_IND_FLAG,
P1.TECHNICAL_INDEX,
P1.SIZE_LITERAL,
P1.[ORIGINAL FCST QTY],
P1.WITHDRAWN_QUANTITY,
P1.[REM PLAN QTY],
P2.FUTURE_FCST,
SUM(I1.ON_HAND_QUANTITY) AS 'TOTAL_OH_INV'
FROM
(SELECT PLANT_CODE,
STOCK_CATEGORY,
MATERIAL,
DISTRIBUTION_VERSION_CODE,
PERIOD_CODE,
REQUIREMENTS_DATE,
PLANNED_QTY AS 'REM PLAN QTY',
VERSION_IND_FLAG,
SIZE_LITERAL,
WITHDRAWN_QUANTITY,
TECHNICAL_INDEX,
PLANNED_QTY + WITHDRAWN_QUANTITY AS 'ORIGINAL FCST QTY'
FROM VW_PLANNED_REQMNTS_TXT
WHERE PLANT_CODE IN ('6040','6041')
AND STOCK_CATEGORY IN ('A60385000','A60385003')
AND DISTRIBUTION_VERSION_CODE IN ('00','01','ZU','Z2')
AND REQUIREMENTS_DATE < GETDATE() - @daynumber
AND PLANNED_QTY > 0) AS P1
LEFT OUTER JOIN
(SELECT
SUM(PLANNED_QTY) AS 'FUTURE_FCST',
MATERIAL,
TECHNICAL_INDEX
FROM VW_PLANNED_REQMNTS_TXT P
WHERE REQUIREMENTS_DATE >= GETDATE()
AND PLANNED_QTY > 0
AND STOCK_CATEGORY IN ('A60385000','A60385003')
GROUP BY MATERIAL,
TECHNICAL_INDEX,
SIZE_LITERAL) AS P2
ON P1.MATERIAL = P2.MATERIAL
AND P1.TECHNICAL_INDEX = P2.TECHNICAL_INDEX
LEFT OUTER JOIN
(SELECT ON_HAND_QUANTITY,
TECHNICAL_INDEX,
MATERIAL,
STOCK_CATEGORY,
PLANT_CODE
FROM VW_INVENTORY I
WHERE STOCK_CATEGORY IN ('A60385000','A60385003')
GROUP BY TECHNICAL_INDEX,
MATERIAL,
STOCK_CATEGORY,
ON_HAND_QUANTITY,
PLANT_CODE) AS I1
ON P1.MATERIAL = I1.MATERIAL
AND P1.TECHNICAL_INDEX = I1.TECHNICAL_INDEX
AND P1.PLANT_CODE = I1.PLANT_CODE
LEFT OUTER JOIN
(SELECT M.material_number,
M.material_desc,
M.division_id,
D.division_desc,
M.gender_id,
G.gender_desc
FROM MaterialMaster.dbo.Material M
JOIN MaterialMaster.dbo.Ref_Division D
ON M.division_id = D.division_id
JOIN MaterialMaster.dbo.Ref_Gender G
ON M.gender_id = G.gender_id) AS M1
ON P1.MATERIAL = M1.material_number
GROUP BY P1.PLANT_CODE,
P1.STOCK_CATEGORY,
P1.MATERIAL,
P1.DISTRIBUTION_VERSION_CODE,
P1.PERIOD_CODE,
P1.REQUIREMENTS_DATE,
P1.[REM PLAN QTY],
P1.VERSION_IND_FLAG,
P1.WITHDRAWN_QUANTITY,
P1.SIZE_LITERAL,
P1.TECHNICAL_INDEX,
P1.[ORIGINAL FCST QTY],
P2.FUTURE_FCST,
M1.material_desc,
M1.division_id,
M1.division_desc,
M1.gender_id,
M1.gender_desc) AS T1 /* T1 Represents "Total" for all */
GROUP BY PLANT_CODE,
STOCK_CATEGORY,
MATERIAL,
material_desc,
division_id,
division_desc,
gender_id,
gender_desc,
DISTRIBUTION_VERSION_CODE,
PERIOD_CODE,
REQUIREMENTS_DATE,
VERSION_IND_FLAG,
TECHNICAL_INDEX,
SIZE_LITERAL,
[ORIGINAL FCST QTY],
WITHDRAWN_QUANTITY,
[REM PLAN QTY],
FUTURE_FCST,
TOTAL_OH_INV
我有一条经验法则。如果我必须使用“选择不同”，我做了一些错误的事情，或者至少我可以做得更好。 [本文]（http://weblogs.sqlteam.com/markc/archive/2008/11/11/60752.aspx）解释得很好。 – mason 2014-09-13 03:13:02
简而言之 - 从未 - 需要在已使用“group by”的单个查询中包含“distinct”;它只是多余的。 @mason我也喜欢那篇文章 – 2014-09-13 07:55:47
'select distinct'产生**行**，它们是唯一的，**整行**被考虑。所以，如果你已经使用选择不同的，你仍然认为信息不是“独特的”，那么你只是没有使用正确的技术。但是：没有你通过一些数据展示你的问题真的是我们可能猜测了很长时间;但通常row_number（）解决了这种情况。您确实需要提供样本数据和预期结果（或更好地设置sqlfiddle） – 2014-09-13 08:01:51