表:FooData
ID Count 1 54 2 42 3 33 4 25 5 16 6 9 8 5 9 3 10 2
我想获取Count列的总和大于某个百分比的记录,如90.
对于这个例子,所有记录的总和是189,其中90%是170.1,我们可以看到ID 1到6的总和是179,它大于170.1所以记录ID 6应该是返回。
顺便说一下,不允许使用临时表,因为我需要在函数中执行它。
答案 0 :(得分:3)
三角形连接的另一个版本。
declare @T table(ID int primary key, [Count] int)
insert into @T values (1, 54), (2, 42), (3, 33),(4, 25), (5, 16), (6, 9), (8, 5), (9, 3),(10, 2)
;with R(ID, [Count], Running) as
(
select T1.ID,
T1.[Count],
cast(T3.[Count] as float)
from @T as T1
cross apply (select sum(T2.[Count])
from @T as T2
where T1.ID >= T2.ID) as T3([Count])
),
T(Total) as
(
select sum([Count])
from @T
)
select top 1 R.ID, R.[Count], R.Running
from R
inner join T
on R.Running / T.Total > 0.9
order by R.ID
答案 1 :(得分:1)
试试这个:
SELECT TOP 1
t2.id,
SUM(t1.value) AS runningTotal
FROM FooData t1
INNER JOIN FooData t2
ON t1.id <= t2.id
GROUP BY t2.id
HAVING SUM(t1.value) * 100. /
(SELECT SUM(value) FROM @FooData) > 90
ORDER BY SUM(t1.value)
但请注意triangular joins in running totals的潜在性能问题。