问题
I have an SQL Server table of the following structure:
id TransDate PartType
======================================
1 2016-06-29 10:23:00 A1
2 2016-06-29 10:30:00 A1
3 2016-06-29 10:32:00 A2
4 2016-06-29 10:33:00 A2
5 2016-06-29 10:35:00 A2
6 2016-06-29 10:39:00 A3
7 2016-06-29 10:41:00 A4
I need a SELECT
statement that will output a table that finds the changes in PartType
, and that looks like this (for SSRS purposes):
PartType StartTime EndTime
=======================================================
A1 2016-06-29 10:23:00 2016-06-29 10:32:00
A2 2016-06-29 10:32:00 2016-06-29 10:39:00
A3 2016-06-29 10:39:00 2016-06-29 10:41:00
A4 2016-06-29 10:41:00 NULL
Note that the StartTime
always picks up from the last EndTime
, unless it's the first record in the table.
What should my SELECT
statement be? I can't seem to get the intended results.
EDIT: I'm using SQL Server 2008 R2; I should've specified that.
回答1:
With SQL Server 2012 and later, you can use this:
declare @t table (id int, transdate datetime2(0), parttype char(2))
insert @t
values
(1, '2016-06-29 10:23:00', 'A1'),
(2, '2016-06-29 10:30:00', 'A1'),
(3, '2016-06-29 10:32:00', 'A2'),
(4, '2016-06-29 10:33:00', 'A2'),
(5, '2016-06-29 10:35:00', 'A2'),
(6, '2016-06-29 10:39:00', 'A3'),
(7, '2016-06-29 10:41:00', 'A4')
;with x as (
select *, row_number() over(partition by parttype order by transdate) rn
from @t
)
select parttype, transdate starttime, lead(transdate) over (order by transdate) from x where rn = 1
回答2:
Hmmm, here is one method using outer apply
and group by
:
select t1.PartType, min(t1.TransDate) as StartTime, t2.TransDate
from t t1
outer apply
(select top 1 t2.*
from t t2
where t2.PartType <> t1.PartType and t2.TransDate > t1.TransDate
order by t2.TransDate asc
) t2
group by t1.PartType, t2.TransDate;
来源:https://stackoverflow.com/questions/38100062/find-row-changes-and-output-to-table