我正在尝试创建一个weekofyear返回的查询，如果该周有4个或更多DISTINCT(DELIVERYDATE)，则返回值。因此，如果只有3天，它将返回一个空值。如果是4个或更多DISTINCT(DELIVERYDATE)，则返回＆＃39; weekofyear＆＃39;值。
SELECT READYDATE,TOTALAMOUNT,ORDER_TRACK_NUMBER,
(CONVERT(VARCHAR(4),LEFT(DELIVERYDATE,4)) + '' + CONVERT(VARCHAR(4),DATEPART( wk, DELIVERYDATE))) AS WEEKOfYear
FROM CurrentOrder COrder
WHERE
COrder.CANCELLED = 0
AND COrder.DELIVERYDATE>='20171116'
AND COrder.DELIVERYDATE<='20171116'
ORDER BY DISPID
查询后我想要的预期输出
READYDATE TOTALAMOUNT ORDERNUM DELIVERYDATE WEEKOfYear
2017-11-16 11.24 6784230 2017-11-16 201746
2017-11-16 0.00 6788903 2017-11-16 201746
2017-11-16 91.85 6789075S 2017-11-16 201746
2017-11-16 640.00 6789080S 2017-11-16 201746
2017-11-16 0.00 6789213S 2017-11-16 201746
2017-11-15 0.00 6789098S 2017-11-15 NULL
2017-11-15 0.00 6789103S 2017-11-15 NULL
2017-11-15 0.00 6789110S 2017-11-15 NULL
2017-11-13 250.00 6789751S 2017-11-13 NULL
2017-11-13 250.00 6789756S 2017-11-13 NULL
2017-11-13 25.00 6789773S 2017-11-13 NULL
2017-11-14 242.00 6789779S 2017-11-14 201746
2017-11-14 23.88 6789816S 2017-11-14 201746
2017-11-14 190.00 6789833S 2017-11-14 201746
2017-11-14 215.00 6789845S 2017-11-14 201746
2017-11-14 108.00 6789873S 2017-11-14 201746
CREATE TABLE [dbo].[CurrentOrder](
[OrderId] [int] NOT NULL,
[ORDER_TRACK_NUMBER] [varchar](50) NULL,
[TOTALAMOUNT] [decimal](18, 10) NULL,
[READYDATE] [date] NULL,
[DELIVERYDATE] [date] NULL,
[CANCELLED] [tinyint] NULL,
[DispID] [int] NULL,
PRIMARY KEY CLUSTERED
(
[OrderId] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
GO
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (1, N'6784230', CAST(11.0000000000 AS Decimal(18, 10)), CAST(0x883D0B00 AS Date), CAST(0x883D0B00 AS Date), 0, 56)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (2, N'6788903', CAST(0.0000000000 AS Decimal(18, 10)), CAST(0x883D0B00 AS Date), CAST(0x883D0B00 AS Date), 0, 56)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (3, N'6789075S', CAST(92.0000000000 AS Decimal(18, 10)), CAST(0x883D0B00 AS Date), CAST(0x883D0B00 AS Date), 0, 56)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (4, N'6789080S', CAST(640.0000000000 AS Decimal(18, 10)), CAST(0x883D0B00 AS Date), CAST(0x883D0B00 AS Date), 0, 56)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (5, N'6789213S', CAST(0.0000000000 AS Decimal(18, 10)), CAST(0x883D0B00 AS Date), CAST(0x883D0B00 AS Date), 0, 56)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (6, N'6789098S', CAST(0.0000000000 AS Decimal(18, 10)), CAST(0x873D0B00 AS Date), CAST(0x873D0B00 AS Date), 0, 66)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (7, N'6789103S', CAST(0.0000000000 AS Decimal(18, 10)), CAST(0x873D0B00 AS Date), CAST(0x873D0B00 AS Date), 0, 66)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (8, N'6789110S', CAST(0.0000000000 AS Decimal(18, 10)), CAST(0x873D0B00 AS Date), CAST(0x873D0B00 AS Date), 0, 66)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (9, N'6789751S', CAST(250.0000000000 AS Decimal(18, 10)), CAST(0x853D0B00 AS Date), CAST(0x853D0B00 AS Date), 0, 65)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (10, N'6789756S', CAST(250.0000000000 AS Decimal(18, 10)), CAST(0x853D0B00 AS Date), CAST(0x853D0B00 AS Date), 0, 65)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (11, N'6789773S', CAST(25.0000000000 AS Decimal(18, 10)), CAST(0x853D0B00 AS Date), CAST(0x853D0B00 AS Date), 0, 65)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (12, N'6789779S', CAST(242.0000000000 AS Decimal(18, 10)), CAST(0x863D0B00 AS Date), CAST(0x863D0B00 AS Date), 0, 39)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (13, N'6789816S', CAST(24.0000000000 AS Decimal(18, 10)), CAST(0x863D0B00 AS Date), CAST(0x863D0B00 AS Date), 0, 39)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (14, N'6789833S', CAST(190.0000000000 AS Decimal(18, 10)), CAST(0x863D0B00 AS Date), CAST(0x863D0B00 AS Date), 0, 39)
INSERT [dbo].[CurrentOrder] ([OrderId], [ORDER_TRACK_NUMBER], [TOTALAMOUNT], [READYDATE], [DELIVERYDATE], [CANCELLED], [DispID]) VALUES (15, N'6789845S', CAST(215.0000000000 AS Decimal(18, 10)), CAST(0x863D0B00 AS Date), CAST(0x863D0B00 AS Date), 0, 39)
说明： - DeliveryDate - 2017-11-16超过4，因此会显示WEEKOfYear中的值，但对于DeliveryDate＆2017; 2017-11-15 has only three orders so this is displaying Null same as for the date 2017-11- 13 . But again DELIVERDATE 2017-11-14`在该日期有超过4个订单，因此显示的是WEEKOfYear。
提前感谢您的帮助。
答案 0 :(得分：2)
您可以使用count()和over()子句来获取每个交付日期的行数，以及确定是否显示weekOfYear或null的情况，如下所示：
SELECT READYDATE,
TOTALAMOUNT,
ORDER_TRACK_NUMBER,
CASE WHEN COUNT(*) OVER(PARTITION BY DELIVERYDATE) > 3 THEN
CAST(DATEPART(YEAR, DELIVERYDATE) as char(4)) +
CAST(DATEPART(WEEK, DELIVERYDATE) as char(2))
ELSE
NULL
END As WeekOfYear
FROM CurrentOrder
WHERE CANCELLED = 0
ORDER BY DISPID
结果：
READYDATE TOTALAMOUNT ORDER_TRACK_NUMBER WeekOfYear
2017-11-14 242,0000000000 6789779S 201746
2017-11-14 24,0000000000 6789816S 201746
2017-11-14 190,0000000000 6789833S 201746
2017-11-14 215,0000000000 6789845S 201746
2017-11-16 11,0000000000 6784230 201746
2017-11-16 0,0000000000 6788903 201746
2017-11-16 92,0000000000 6789075S 201746
2017-11-16 640,0000000000 6789080S 201746
2017-11-16 0,0000000000 6789213S 201746
2017-11-13 250,0000000000 6789751S NULL
2017-11-13 250,0000000000 6789756S NULL
2017-11-13 25,0000000000 6789773S NULL
2017-11-15 0,0000000000 6789098S NULL
2017-11-15 0,0000000000 6789103S NULL
2017-11-15 0,0000000000 6789110S NULL
答案 1 :(得分：1)
此方法使用common table expression (CTE)查找包含4个或更多记录的每个交货日期。使用交货日期计算一周中的一周。然后，我们可以使用outer join将新字段附加到原始表中的每个匹配记录。
WITH DWeek AS
(
-- Calculate week of year for every delivery date with 4 or more records.
SELECT
DELIVERYDATE,
(CONVERT(VARCHAR(4),LEFT(DELIVERYDATE,4)) + '' + CONVERT(VARCHAR(4),DATEPART( wk, DELIVERYDATE))) AS WEEKOfYear
FROM
CurrentOrder
WHERE
DELIVERYDATE IS NOT NULL
AND CANCELLED = 0
AND DELIVERYDATE >= '20171114'
AND DELIVERYDATE <= '20171116'
GROUP BY
DELIVERYDATE
HAVING
COUNT(1) >= 4
)
SELECT
co.*,
s.WEEKOfYear
FROM
CurrentOrder AS co
LEFT OUTER JOIN DWeek AS s ON s.DELIVERYDATE = co.DELIVERYDATE
;
返回
OrderId ORDER_TRACK_NUMBER TOTALAMOUNT READYDATE CANCELLED DELIVERYDATE WEEKOfYear
1 6784230 11.2400000000 2017-11-16 0 2017-11-16 201746
2 6788903 0.0000000000 2017-11-16 0 2017-11-16 201746
3 6789075S 91.8500000000 2017-11-16 0 2017-11-16 201746
4 6789080S 640.0000000000 2017-11-16 0 2017-11-16 201746
5 6789213S 0.0000000000 2017-11-16 0 2017-11-16 201746
6 6789098S 0.0000000000 2017-11-15 0 2017-11-15 NULL
7 6789103S 0.0000000000 2017-11-15 0 2017-11-15 NULL
8 6789110S 0.0000000000 2017-11-15 0 2017-11-15 NULL
9 6789751S 250.0000000000 2017-11-13 0 2017-11-13 NULL
10 6789756S 250.0000000000 2017-11-13 0 2017-11-13 NULL
11 6789773S 25.0000000000 2017-11-13 0 2017-11-13 NULL
12 6789779S 242.0000000000 2017-11-14 0 2017-11-14 201746
13 6789816S 23.8800000000 2017-11-14 0 2017-11-14 201746
14 6789833S 190.0000000000 2017-11-14 0 2017-11-14 201746
15 6789845S 215.0000000000 2017-11-14 0 2017-11-14 201746
修改
我的原始答案错过了OP的一些重要细节。这是一个完全重写的答案。
答案 2 :(得分：0)
这将在你想要的结果中提到： -
select mastr.* , case when mastr.deliverydate = del.deliverydate then (CONVERT(VARCHAR(4),LEFT(mastr.DELIVERYDATE,4)) + '' + CONVERT(VARCHAR(4),DATEPART( wk, mastr.DELIVERYDATE))) else null end as 'WeekOfYear' from
[dbo].[CurrentOrder] mastr
left join
(Select distinct deliverydate from
(select deliverydate, rank() over(partition by deliverydate order by orderid) as cnt
from [dbo].[CurrentOrder] where [CANCELLED] = 0) ranked
Where cnt > 3) del
on mastr.deliverydate = del.deliverydate