Question

我有2个日期，StartDate和EndDate：

Declare @StartDate date='2018/01/01', @Enddate date ='2018/12/31'

然后在mytable表中有一些带有date和value的数据：

----------------------------
 ID    date          value
----------------------------
  1  2018/02/14      4
  2  2018/09/26      7
  3  2017/09/20      2

数据可能在2018年之前开始，并且如果它在@startdate之前存在，则在值之前否则得到0 我正在寻找一个看起来像这样的结果：

-----------------------------------
fromdate      todate       value
-----------------------------------
2018/01/01    2018/02/13     2
2018/02/14    2018/09/25     4
2018/09/26    2018/12/31     7

第一个fromdate来自@StartDate，最后一个todate来自@Enddate，并且应该生成其他数据。

我希望在SQL查询中得到它。我使用sql-server 2016

Answer 1

with cte as 
(
    select 0 as row_num, @StartDate as start_date, 0 as val
    UNION
    select ROW_NUMBER() OVER(ORDER BY start_date) as row_num, * from input
)
select curr.start_date 
  , DATEADD(day,-1,ISNULL(nex.start_date,DATEADD(day,1,@Enddate))) as end_date 
  , curr.val 
from cte curr
left join cte nex on curr.row_num = nex.row_num - 1;

您可以在此处找到模拟：https://rextester.com/EIAXW23839

Answer 2

您可以使用CTE创建完整的日期范围，然后使用LEAD创建ToDate列：

DECLARE @FromDate date = '20180101',
        @ToDate date = '20181231';


WITH VTE AS(
    SELECT ID,
           CONVERT(date,[date]) [date], --This is why using keywords for column names is a bad idea
           [value]
    FROM (VALUES(1,'20180214',4),
                (2,'20180926',7),
                (3,'20170314',4))V(ID,[date],[value])),
Dates AS(
    SELECT [date]
    FROM VTE V
    WHERE V.[date] BETWEEN @FromDate and @ToDate
    UNION ALL
    SELECT [date]
    FROM (VALUES(@FromDate))V([date]))
SELECT D.[date] AS FromDate,
       LEAD(DATEADD(DAY, -1,D.[date]),1,@ToDate) OVER (ORDER BY D.[date]) AS ToDate,
       ISNULL(V.[value],0) AS [value]
FROM Dates D
     LEFT JOIN VTE V ON D.[date] = V.[date];

numpy.where

通过一些值在两个范围日期之间生成数据

2 个答案: