以下SQL语句，使用sqlserver数据库

方差的计算公式为： 方差 = [(x1 - 平均数)^2 + (x2 - 平均数)^2 + … + (xn - 平均数)^2] / n 其中 x1、x2、…、xn 是样本中的各个数据，n 是样本数量。

SELECT AVG(valu) AS avg_valu,COUNT(valu) AS cnt,SUM(POWER(avg_valu-valu,2))/COUNT(valu) AS  [方差],AVG(POWER(avg_valu-valu,2)) AS  [方差2],SQRT(AVG(POWER(avg_valu-valu,2))) AS [标准差1],stdevp(valu)  AS [标准差2],stdev(valu)  AS [样本标准差]
FROM (SELECT  valu,AVG(valu)OVER() AS avg_valuFROM (VALUES (10.12),(10.32),(10.22),(10.52),(10.42)) AS out_tmp(valu)
) AS tmp

CPK 是一种用于衡量过程能力的指标， 它考虑了生产过程的均值和标准差与规格限的关系。 CPK（过程能力指数）的计算公式为：

CPK = min{(USL - X̄) / 3σ, (X̄ - LSL) / 3σ} 其中：

• USL 表示规格上限；
• LSL 表示规格下限；
• X̄ 表示样本平均值；
• σ 表示样本标准差。

这个指标用于衡量过程的稳定性和能力。

  SELECT 'MIN((上限-平均值)/(3*标准差), (平均值-下限)/(3*标准差))'  AS 'CPK公式',([上限]-[平均值])/(3*[标准差]) AS [上],([平均值]-[下限])/(3*[标准差]) AS [下],CASE WHEN [标准差] >= 0 AND (([上限]-[平均值])/(3*[标准差])) > (([平均值]-[下限])/(3*[标准差]))THEN  ([平均值]-[下限])/(3*[标准差])ELSE  ([上限]-[平均值])/(3*[标准差]) END AS [cpk]FROM ( SELECT [上限],[下限],[平均值],CASE WHEN 0= [标准差] THEN NULL ELSE [标准差] END  AS [标准差]FROM (SELECT 11 AS [上限],0 AS [下限],avg(aaaa)  	 AS [平均值],stdevp(aaaa) AS [标准差],stdev(aaaa)  AS [样本标准差]FROM (VALUES (10.022),(10.3352),(10.278),(10.561),(10.630)) as out_tmp(aaaa)) AS in_tmp)as tmp