同比：
SELECT DISTINCT
B.dateM,B.dateY,B.Total,B.Total+B.Total*AVG( B.per ) OVER ( PARTITION BY B.dateM ) as predictionTotal
, AVG( B.per ) OVER ( PARTITION BY B.dateM ) AS [AvgPer]
,MAX( B.per ) OVER ( PARTITION BY B.dateM ) AS [MaxPer]
,Min( B.per ) OVER ( PARTITION BY B.dateM ) AS [MinPer] from
(SELECT A.dateM,A.dateY
,A.[Total] - A.[PreviousTotal] as GrowthTotal,(A.[Total] - A.[PreviousTotal])/A.[PreviousTotal] as per,A.[Total]
, A.[Month]
FROM
(
select s.dateM,dateY,COUNT(*) OVER ( PARTITION BY s.dateM ) AS [Month],s.[Total] AS [Total]
, LAG( s.[Total] ) OVER ( PARTITION BY s.dateM ORDER BY s.dateY ) AS [PreviousTotal]
from (SELECT datepart(mm,[month]) as dateM, datepart(yyyy,[month]) as dateY,sum([Total]) Total
FROM [KS_DataBase3.0_QJ].[dbo].[KS_MonthVbt] where [Month]>='2017' and [Month]<'2019'
GROUP BY [month]) s
) AS A
) B
环比计算比较简单
SELECT o.*
FROM (SELECT o.*,
(total / LAG(total) OVER (ORDER BY month) - 1) as growth
FROM [KS_DataBase3.0_QJ].[dbo].[KS_MonthVbt] o
) o
WHERE [Month] <= '2019-10-01'
ORDER BY [Month];