问题描述
表StatusHistory
包含:
ID Status DtSince DtUntil
2415891 0 20200816 20200917
2415891 4 20200917 NULL
2415892 0 20200904 20200905
2415892 4 20200905 NULL
我需要为每个ID选择每个月初和月末的状态。 用户将输入所需的日期。例如:
用户输入-> @Month = '202009'
select语句必须返回dates 20200901 and 20201001
的状态
对于新记录并具有dtSince> 20200901的记录,它应返回活动行。例如,对于2415892
,它应返回4
。
因此@Month = 202009的结果集应为:
Id BeginningOfMonth_status EndOfMonth_status
2415891 0 4
2415892 4 4
这是我的解决方法
select * from
(
select * from
(
select *,ROW_NUMBER() over (partition by id order by dtSince desc) r from
dbo.StatusHistory
where DtUntil is null or (DtUntil <20201001and dtSince>=20200901) --2432290
)x
where x.r=1
)s1
inner join
(
select * from
(
select *,ROW_NUMBER() over (partition by id order by dtSince desc) r from
dbo.StatusHistory
where DtUntil is null or (DtUntil <20201001and dtSince>=20200901) --2432290
)x
where x.r=2
)s2
on s1.id = s2.id
问题:
1)有没有比我的代码复杂的东西了?
2)以上所述,我无法满足For records that are new and have dtSince>20200901
的要求
解决方法
嗯。 。 。您可以使用where
子句获取涵盖特定月份的所有行。然后使用row_number()
枚举行和条件聚合:
select id,max(case when seqnum_asc = 1 then status end) as beginning_of_month,max(case when seqnum_desc = 1 then status end) as end_of_month
from (select t.*,row_nunber() over (partition by id order by dtsince) as seqnum_asc,row_nunber() over (partition by id order by dtsince desc) as seqnum_desc
from t
where dtsince < dateadd(month,1,convert(date,@Month + '01'))) and
(dtuntil >= convert(date,@Month + '01') or dtuntil is null)
) t
group by id