SQL SERVER 2012：每人行数的标准差

Question

我有 table ITEMS:

Itemid, 
ItemName, 
OwnerID

每一行代表一个项目。一个所有者可以有多个项目。我想计算每个所有者 items/rows 的标准差。是否可以使用 STDEV 内置函数来计算？

我试过了，但似乎不行：

select STDEV(COUNT(*)) from dbo.ITEMS 
GROUP BY ITEMS.OwnerID

Answer 1

我觉得

    SELECT STDEV(IQ.cnt) as SDofCounts from --'from' was missing
        (select ITEMS.OwnerID, COUNT(0) as cnt from dbo.ITEMS 
            GROUP BY ITEMS.OwnerID
        ) IQ

++++++++++++++++++

如果你运行这个要看一些数据

        select ITEMS.OwnerID, COUNT(0) as cnt, COUNT(0) * COUNT(0) as EXS   from dbo.ITEMS 
            GROUP BY ITEMS.OwnerID

那你可以做

SELECT STDEV(IQ.cnt) as SDofCounts, 
       SUM(EXS) as SigmaEXsquare, 
       SUM(1.0 * cnt) / COUNT(0) as MU, 
       SUM(EXS) - (SUM(1.0 * cnt) / COUNT(0)) * (SUM(1.0 * cnt) / COUNT(0)) as Variance, 
       SQRT(SUM(EXS) - (SUM(1.0 * cnt) / COUNT(0)) * (SUM(1.0 * cnt) / COUNT(0))) AS SDcalc
        from
        (select ITEMS.OwnerID, COUNT(0) as cnt, COUNT(0) * COUNT(0) as EXS from dbo.ITEMS 
            GROUP BY ITEMS.OwnerID
        ) IQ

SQL SERVER 2012：每人行数的标准差

SQL SERVER 2012: Standard Deviation of rows per person

sql

sql-server

sql-server-2012

standard-deviation

stdev