SQL：按连续日期计算记录数，即使某个日期不存在记录

Question

我使用的是 SQL Server 2008 R2。

我正在查询医院预约空档表，并尝试返回一个列表，其中列出了标记为已预订的特定医生的预约空档数量，按周数/年分组。

有些情况下几周还没有任何预约，但我希望结果列出所有未来几周，即使预约的预约空档数为零。

我正在寻找的输出是这样的：

-------------------------------------------
Year | Week Number | Number of Booked Slots
-------------------------------------------
2017 |     48      |        10            
2017 |     49      |         0
2017 |     50      |         4
2017 |     51      |         2
2017 |     52      |         0
2018 |      1      |         5

我知道，聚合结果的标准选择不会显示记录数为零的那几周，因为没有什么可以返回 - 所以我试图通过使用 cte 首先生成一个列表来解决这个问题未来几周。

但是，尽我所能，我无法让查询显示零周...

我已经看到了许多类似问题的解决方案，但尽管进行了实验，但我无法将它们应用于我的特定问题（包括 SQL Count to include zero values）

这是迄今为止我编写的查询的最新迭代。

    WITH CTE_Dates AS
    (
    SELECT DISTINCT Slot_Start_Date AS cte_date
    FROM [Outpatients.vw_OP_Clinic_Slots WITH (NOLOCK)
    )

SELECT 
    DATEPART(year,OPCS.Slot_Start_Date) [Year]
    ,DATEPART(week,OPCS.Slot_Start_Date) [Week Number]
    ,count(OPCS.Slot_Start_Date) [Number of Booked Slots]

FROM 
    Outpatients.vw_OP_Clinic_Slots OPCS WITH (NOLOCK)
    LEFT OUTER JOIN CTE_Dates ON OPCS.Slot_Start_Date=CTE_Dates.cte_date
    LEFT OUTER JOIN Outpatients.vw_OP_Clinics CLIN ON OPCS.Clinic_Code=CLIN.Clinic_Code

WHERE
    OPCS.Slot_Start_Date >= '14/08/2017'
    AND OPCS.Booked_Flag = 'Y'
    AND CLIN.Lead_Healthcare_Professional_Name = 'Dr X'

GROUP BY
    DATEPART(year,OPCS.Slot_Start_Date)
    ,DATEPART(week,OPCS.Slot_Start_Date)

ORDER BY 
    DATEPART(year,OPCS.Slot_Start_Date)asc
    ,DATEPART(week,OPCS.Slot_Start_Date)asc

它返回给我的结果是正确的，但我只需要它在列表中包含计数为零的那些周。

请谁能解释我哪里出错了？我猜我没有正确加入 cte，但我已经尝试了产生相同结果的右连接和左连接。我还尝试通过交换上述查询语句和 cte 来反转查询，但这也不起作用。

感谢任何人可以提供的任何指导。

Answer 1

只需 RIGHT JOIN @ListOfWeeks 表到您拥有的结果集：

DECLARE @ListOfWeeks TABLE ([Week_No] int, [Year_Number] int);

DECLARE @i tinyint = 1, @y int = 2010;

WHILE @i <= 52 AND @y < 2018
BEGIN
    INSERT INTO @ListOfWeeks([Week_No], [Year_Number]) VALUES (@i, @y);

    IF @i = 52 BEGIN
       SET @i = 0
       SET @y +=1
       END
     SET @i += 1
END

SELECT * FROM @ListOfWeeks

WITH [Your_Part] AS(
SELECT 
    DATEPART(year,OPCS.Slot_Start_Date) [Year]
    ,DATEPART(week,OPCS.Slot_Start_Date) [Week Number]
    ,count(OPCS.Slot_Start_Date) [Number of Booked Slots]

FROM 
    Outpatients.vw_OP_Clinic_Slots OPCS WITH (NOLOCK)
    LEFT OUTER JOIN CTE_Dates ON OPCS.Slot_Start_Date=CTE_Dates.cte_date
    LEFT OUTER JOIN Outpatients.vw_OP_Clinics CLIN ON OPCS.Clinic_Code=CLIN.Clinic_Code

WHERE
    OPCS.Slot_Start_Date >= '14/08/2017'
    AND OPCS.Booked_Flag = 'Y'
    AND CLIN.Lead_Healthcare_Professional_Name = 'Dr X'

GROUP BY
    DATEPART(year,OPCS.Slot_Start_Date)
    ,DATEPART(week,OPCS.Slot_Start_Date)
),
SELECT xxx.[Year_Number], xxx.[Week_No], yp.[Number of Booked Slots]
FROM [Your_Part] yp
RIGHT JOIN @ListOfWeeks xxx ON  yp.[Year] = xxx.[Year_Number] AND yp.[Week Number] = xxx.[Week_No]

Answer 2

IF OBJECT_ID(N'tempdb..##cal_weeks_temp', N'U') IS NOT NULL   
DROP TABLE ##cal_weeks_temp;  
create table ##cal_weeks_temp (date_of_week date, week_num int)
declare @start_date date 
declare @end_date date 
set @start_date='01/01/2017'
set @end_date='12/31/2018'
while @start_date<@end_date
begin
    set @start_date=dateadd(day,1,@start_date)  
    insert into ##cal_weeks_temp values (@start_date,DATEPART(week,@start_date))
end
select YEAR(t1.date_of_week) 'YEAR',t1.week_num,
sum(case convert(varchar,t2.BookedTime,105) when convert(varchar,t1.date_of_week,105)  then 1 else 0 end) 'count'
 from ##cal_weeks_temp t1
left join  Your_Table t2
on convert(varchar,t2.BookedTime,105)=convert(varchar,t1.date_of_week,105) 
group by YEAR(t1.date_of_week) ,t1.week_num
order by YEAR(t1.date_of_week) ,t1.week_num