我有这张桌子:
[SDate - DateTime] [Value] - Represents a boolean
2010-11-16 10:10:00 1
2010-11-16 10:11:00 0
2010-11-16 10:12:00 1
2010-11-16 10:13:00 1
我需要一个查询来计算值变化的次数。
例如:
2010-11-16 10:10:00 0
2010-11-16 10:11:00 1
2010-11-16 10:12:00 0
2010-11-16 10:13:00 0
2010-11-16 10:14:00 1
2010-11-16 10:15:00 0
2010-11-16 10:16:00 1
...
5 changes
2010-11-16 10:10:00 0
2010-11-16 10:11:00 0
2010-11-16 10:12:00 0
2010-11-16 10:13:00 1
2010-11-16 10:14:00 1
2010-11-16 10:15:00 1
2010-11-16 10:16:00 1
...
1 change
答案 0 :(得分:8)
您可以使用lag()
:
select count(*)
from (select t.*, lag(value) order by (sdate) as prev_value
from table t
) t
where prev_value <> value ;
答案 1 :(得分:2)
这也适用于早期版本..
;WITH cte
AS (SELECT Row_number()OVER(ORDER BY sdates) rn,*
FROM <tablename>)
SELECT Sum(CASE WHEN a.boolvalue = b.boolvalue THEN 0 ELSE 1 END)
FROM cte a
JOIN cte b
ON a.rn = b.rn + 1
答案 2 :(得分:1)
Pre-SQL Server 2012解决方案:
declare @t table ( sdate datetime, value int )
insert into @t(sdate, value)
values
--('2010-11-16 10:10:00', 0),
--('2010-11-16 10:11:00', 1),
--('2010-11-16 10:12:00', 0),
--('2010-11-16 10:13:00', 0),
--('2010-11-16 10:14:00', 1),
--('2010-11-16 10:15:00', 0),
--('2010-11-16 10:16:00', 1)
('2010-11-16 10:10:00', 0),
('2010-11-16 10:11:00', 0),
('2010-11-16 10:12:00', 0),
('2010-11-16 10:13:00', 1),
('2010-11-16 10:14:00', 1),
('2010-11-16 10:15:00', 1),
('2010-11-16 10:16:00', 1)
;with t_with_seq as (
select
t.*,
ROW_NUMBER() OVER(ORDER BY t.sdate asc) as seq
from @t t
)
select
COUNT(*)
from t_with_seq r
inner join t_with_seq r_next on r_next.seq = r.seq + 1
where r.value <> r_next.value