SQL - 在日期速率内计算 itemactive 的数量

Question

我有一个 Resources、Projects、StartDate 和 EndDate 的数据集。 每个资源可以被多个项目使用。

我想计算每个季度使用资源的项目数量。

因此，如果项目在特定年份的 Q1 开始并在该年的 Q3 结束，并且 project2 在 Q2 开始并在 Q3 结束，我想计算 Q2 的 2 个项目，因为在 Q1 期间 project1 和 project2 都是活跃。

这是我的数据集：

create table Projects
(Resource_Name varchar(20)
,Project_Name varchar(20)
,StartDate varchar(20)
,EndDate varchar(20)
)


insert into Projects values('Resource 1','Project A','15/01/2013','1/11/2014')
insert into Projects values('Resource 1','Project B','1/03/2013','1/09/2016')
insert into Projects values('Resource 1','Project C','1/04/2013','1/09/2015')
insert into Projects values('Resource 1','Project D','1/06/2013','1/03/2016')
insert into Projects values('Resource 1','Project E','15/01/2013','1/09/2015')
insert into Projects values('Resource 1','Project F','3/06/2013','1/11/2015')

这是我正在寻找的结果：

Resource Name| Year | Quarter|Active Projects
Resource 1     2013     1           2
Resource 1     2013     2           6

Answer 1

您可以通过确定项目处于活动状态的第一季度和最后一个季度，然后使用累积总和来做到这一点。在 SQL Server 2012+ 中，这看起来像

select resource_name, yyyyq, 
       (sum(sum(s)) over (partition by resource_name order by yyyyq) -
        sum(sum(e)) over (partition by resource_name order by yyyyq) +
        e
       ) as activeProjects
from ((select resource_name, datepart(year, startdate) + datepart(quarter, startdate) as yyyyq, 1 as s, 0 as e
       from projects
      ) union all
      (select resource_name, datepart(year, enddate) + datepart(quarter, enddate), 0 as s, 1 as e
       from projects
      )
     ) yq
group by resource_name, yyyyq;

在早期版本中，您可以使用cross apply 执行类似的操作。

Answer 2

使用计数表：

使用Projects 中的日期，生成所有季度及其开始日期和结束日期的列表，在本例中为CteQuarter(sd, ed)。之后，您只需将JOIN Projects 表发送到CteQuarter 以获取重叠日期。最后，GROUP BY 使用日期的YEAR 和Quarter 部分。

SQL Fiddle

WITH CteYear(yr) AS(
    SELECT number
    FROM master..spt_values
    WHERE 
        type = 'P'
        AND number >= (SELECT MIN(YEAR(CONVERT(DATE, StartDate, 103))) FROM Projects)
        AND number <= (SELECT MAX(YEAR(CONVERT(DATE, EndDate, 103))) FROM Projects)
),
CteQuarter(sd, ed) AS(
    SELECT
        DATEADD(QUARTER, q.n - 1, DATEADD(YEAR, cy.yr - 1900, 0)),
        DATEADD(DAY, -1, DATEADD(QUARTER, q.n, DATEADD(YEAR, cy.yr - 1900, 0)))
    FROM CteYear AS cy
    CROSS JOIN(
        SELECT 1 UNION ALL SELECT 2 UNION ALL SELECT 3 UNION ALL SELECT 4
    ) AS q(n)
)
SELECT 
    p.Resource_Name,
    [Year] = DATEPART(YEAR, q.sd),
    [Quarter] = DATEPART(QUARTER, q.sd),
    [Active Projects] = COUNT(*)
FROM Projects p
INNER JOIN CteQuarter q
    ON CONVERT(DATE, StartDate, 103) <= q.ed
    AND CONVERT(DATE, EndDate, 103) >= q.sd
GROUP BY 
    p.Resource_Name,
    DATEPART(YEAR, q.sd),
    DATEPART(QUARTER, q.sd)
ORDER BY
    p.Resource_Name,
    DATEPART(YEAR, q.sd),
    DATEPART(QUARTER, q.sd)

注意事项：

1. Here 是检查重叠日期的好方法。
2. Some common date routines

---

结果：

| Resource_Name | Year | Quarter | Active Projects |
|---------------|------|---------|-----------------|
|    Resource 1 | 2013 |       1 |               3 |
|    Resource 1 | 2013 |       2 |               6 |
|    Resource 1 | 2013 |       3 |               6 |
|    Resource 1 | 2013 |       4 |               6 |
|    Resource 1 | 2014 |       1 |               6 |
|    Resource 1 | 2014 |       2 |               6 |
|    Resource 1 | 2014 |       3 |               6 |
|    Resource 1 | 2014 |       4 |               6 |
|    Resource 1 | 2015 |       1 |               5 |
|    Resource 1 | 2015 |       2 |               5 |
|    Resource 1 | 2015 |       3 |               5 |
|    Resource 1 | 2015 |       4 |               3 |
|    Resource 1 | 2016 |       1 |               2 |
|    Resource 1 | 2016 |       2 |               1 |
|    Resource 1 | 2016 |       3 |               1 |