汇总SQL查询中的拆分范围

时间:2022-09-26 15:47:10

I have a table which contains my server status

我有一个包含我的服务器状态的表

create table ServerStatus
(
    ServerId int, 
    StartTime datetime, 
    Seconds int,
    [State] char(8) 
)

I would like a query that given a start and end date will summarize the time the server spends in each state during that time. I would also like the query to return the amount of time the servers spend in an unknown state.

我想要一个查询,给出一个开始和结束日期将总结服务器在那段时间内在每个州花费的时间。我还希望查询返回服务器在未知状态下花费的时间。

So, for example for the following data

因此,例如对于以下数据

1   2008-01-01 00:00:00.000 120 broken  
1   2008-01-02 00:00:00.000 120 off     
1   2008-01-03 00:00:00.000 240 burning 
1   2008-01-04 00:00:00.000 60  off     
1   2008-01-05 00:00:00.000 60  off     
2   2008-01-01 00:00:00.000 60  broken  
2   2008-01-02 00:00:00.000 30  off     
2   2008-01-03 00:00:00.000 20  burning 
2   2008-01-04 00:00:00.000 600 off     
3   2007-01-04 00:00:00.000 600 off     
4   2007-12-12 00:00:00.000 999999999   onfire  

Provided the range.

提供范围。

select  @start = dateadd(second, 60, '2008-01-01'), 
@fin = dateadd(second, 60, '2008-01-04')

I would like to return the results:

我想返回结果:

1   broken      60
1   burning     240
1   off         180
1   unknown     258720
2   burning     20
2   off         90
2   unknown     259090
4   onfire      259200

This question is somewhat related to: Combining split date ranges in a SQL query

此问题与以下内容有关:在SQL查询中组合拆分日期范围

1 个解决方案

#1


3  

This is the best I have come up with so far:

这是迄今为止我提出的最好的:

declare @start datetime
declare @fin datetime 

select  @start = dateadd(second, 60, '2008-01-01'), @fin = dateadd(second, 60, '2008-01-04')

select ServerId, State, Total = sum(
    case when StartTime > @start and DateAdd(second, Seconds, StartTime) <= @fin
        then Seconds
    when StartTime < @start and DateAdd(second, Seconds, StartTime) <= @fin
        then DateDiff(second, @start, DateAdd(second, Seconds, StartTime)) 
    when StartTime < @start and DateAdd(second, Seconds, StartTime) >= @fin
        then DateDiff(second, @start, @fin)
    else 
        DateDiff(second,StartTime,@fin)
    end)
into #t
from ServerStatus 
where @start < dateadd(second, Seconds, StartTime) 
and @fin > StartTime
group by ServerId, State

insert #t
select ServerId, 'unknown', DateDiff(second, @start, @fin) - sum(Total) 
from #t 
group by ServerId
having DateDiff(second, @start, @fin) > sum(Total)

select * from #t 
order by ServerId, State

#1


3  

This is the best I have come up with so far:

这是迄今为止我提出的最好的:

declare @start datetime
declare @fin datetime 

select  @start = dateadd(second, 60, '2008-01-01'), @fin = dateadd(second, 60, '2008-01-04')

select ServerId, State, Total = sum(
    case when StartTime > @start and DateAdd(second, Seconds, StartTime) <= @fin
        then Seconds
    when StartTime < @start and DateAdd(second, Seconds, StartTime) <= @fin
        then DateDiff(second, @start, DateAdd(second, Seconds, StartTime)) 
    when StartTime < @start and DateAdd(second, Seconds, StartTime) >= @fin
        then DateDiff(second, @start, @fin)
    else 
        DateDiff(second,StartTime,@fin)
    end)
into #t
from ServerStatus 
where @start < dateadd(second, Seconds, StartTime) 
and @fin > StartTime
group by ServerId, State

insert #t
select ServerId, 'unknown', DateDiff(second, @start, @fin) - sum(Total) 
from #t 
group by ServerId
having DateDiff(second, @start, @fin) > sum(Total)

select * from #t 
order by ServerId, State