T-SQL：不使用临时表的日期列表

Question

I have a table

|Start Date|End Date |Value|Avgerage Value Per Day|

|2011-01-01 |2012-01-01| 730 | 2|

I want to turn this table into a View

|Date| Average Value |

2011-01-01 | 2

2011-01-02 | 2

2011-01-03 | 2

.....

2011-12-31 | 2

Is is possible without using temp table to generate a list of date? Any ideas?

Edit

Thanks both of the answers

With recursive view is similar as temp table

I do worry about the performance in a view, caz the view later will be involved in other processes.

I'll try recursive view then, if it doesn't fit, I may just use an hard code date list table .

Answer 1

declare @start datetime
SET @start = '20110501'
declare @end datetime 
SET @end ='20120501'

;with months (date)
AS
(
    SELECT @start
    UNION ALL
    SELECT DATEADD(day,1,date)
    from months
    where DATEADD(day,1,date)<=@end
)
select * from months OPTION (MAXRECURSION 0); 

etc..etc..etc..

Answer 2

Yes you can. This generates the days from the input set and then gives you the ranges you need

Though this technically internally is like temp tables you can create a recursive view :

Create View TestView as
    with Data As -- Pretends to be your table of data replace with  select from your tables
    (
        select Cast('2012-05-01' as DATETIME) [Start Date], Cast('2012-05-02' as DATETIME)  [End Date], 2  [Avgerage Value Per Day]
        union all
        select Cast('2012-04-01' as DATETIME) [Start Date], Cast('2012-04-05' as DATETIME) [End Date], 3  [Avgerage Value Per Day]
    )
    ,AllDates as -- generates all days
    (
         select cast('1900-01-01' as datetime) TheDate
         union all
         select TheDate + 1
         from    AllDates   
         where   TheDate + 1 < '2050-12-31'
    )
    select TheDate [Date], o.[Avgerage Value Per Day]
    from    AllDates
    join Data o on TheDate Between o.[Start Date]   AND   o.[End Date];

you can the query it but you need to ensure that you specify a recursion limit

select * from TestView
OPTION (MAXRECURSION 0)

this gives this result

Date                        Avgerage Value Per Day
2012-04-01 00:00:00.000 3
2012-04-02 00:00:00.000 3
2012-04-03 00:00:00.000 3
2012-04-04 00:00:00.000 3
2012-04-05 00:00:00.000 3
2012-05-01 00:00:00.000 2
2012-05-02 00:00:00.000 2

You can see that from the test data I wanted May 1-2 and April 1-5