使用SQL的案例分析

69次阅读

共计 3691 个字符，预计需要花费 10 分钟才能阅读完成。

丸趣 TV 小编给大家分享一下使用 SQL 的案例分析，相信大部分人都还不怎么了解，因此分享这篇文章给大家参考一下，希望大家阅读完这篇文章后大有收获，下面让我们一起去了解一下吧！

需求

所有部门汇总的结果的打分。大部分指标是根据部门汇总 (SUM 或 AVG) 结果打分。
但面谈率和前 10% 加班平均小时数俩指标，不是根据所有部门 (TEAM) 汇总 (SUM 或 AVG) 结果打分。而是需要根据每个部门的
数据做排序或累加后的结果打分
面谈率计算方式：TEAM1 中季度 1 月份，面谈人数比例目标是 33% -》目标面谈人数 = 部门人数 *33%
实际面谈人数 = 10
TEAM1 中季度 2 月份，面谈人数比例目标是 60% -》目标面谈人数 = 部门人数 *60%
实际面谈人数 = 2 + 10（1 月份实际面谈人数）
TEAM1 中季度 3 月份，面谈人数比例目标是 100% -》目标面谈人数 = 部门人数 *100%
实际面谈人数 = 1 + 12（2 月份实际面谈人数）
这里部门人数是按季度统计每个季度部门的总人数 (且不算领导，在之前存储过程中部门人数已确定，CVS_DEPT_PAX_CNT 表示），因为每季度部门人数或部门关系都可能变更
计算每月实际面谈率 = 实际面谈人数 / CAST(目标面谈人数 AS NUMERIC(19,2)) * 100
注意这里是把整数转换成小数保证不用整除计算。要求是小数
具体 SQL 如下
/* 每个 group 的统计 */

DECLARE @CVS_SING_STAT AS TABLE
(
TEAM_CD VARCHAR(20)
,GRP_CD VARCHAR(20)
,CHECK_RADIO numeric(19,2)

)
INSERT INTO @CVS_SING_STAT
SELECT T.TEAM_CD,T.DEPT_CD GRP_CD,T.CHECK_RADIO
FROM /* 如下内层查询是在 SQL SERVER 2008 上做每行的累积运算 */
SELECT SUBSTRING(BASE.DHSTC_CD,0,5) DH_YEAR ,BASE.STC_MONTH ,SUBSTRING(BASE.DHSTC_CD,5,8) DH_MONTH
,BASE.TEAM_CD,BASE.DEPT_CD,BASE.H09_CNT,BASE.CVS_DEPT_PAX_CNT
,SUM(ROLL.H09_CNT) ACC
,SUM(ROLL.H09_CNT)/BASE.CVS_DEPT_PAX_CNT CVS_RAT
–,33.0/100.0
–,CAST(33.0/100.0 AS NUMERIC(19,2))
–,SUM(ROLL.H09_CNT)/(BASE.CVS_DEPT_PAX_CNT*(33.0/100.0)) 就是实际面谈人数 / 目标面谈人数
–,CAST(SUM(ROLL.H09_CNT)/(BASE.CVS_DEPT_PAX_CNT*(33.0/100.0)) AS NUMERIC(19,2)) FLOATTEST
,CASE WHEN BASE.STC_MONTH = 01 THEN CAST(SUM(ROLL.H09_CNT)/(BASE.CVS_DEPT_PAX_CNT*(33.0/100.0)) AS NUMERIC(19,2))
WHEN BASE.STC_MONTH = 02 THEN CAST(SUM(ROLL.H09_CNT)/(BASE.CVS_DEPT_PAX_CNT*(66.0/100.0)) AS NUMERIC(19,2))
WHEN BASE.STC_MONTH = 03 THEN CAST(SUM(ROLL.H09_CNT)/(BASE.CVS_DEPT_PAX_CNT*(100.0/100.0)) AS NUMERIC(19,2))
END CHECK_RADIO
FROM T_DM_DHSTC_DEPT BASE
,T_DM_DHSTC_DEPT ROLL
WHERE SUBSTRING(BASE.DHSTC_CD,0,5) = SUBSTRING(ROLL.DHSTC_CD,0,5)–YEAR
AND BASE.STC_QUATR = ROLL.STC_QUATR
AND BASE.STC_MONTH = ROLL.STC_MONTH
AND BASE.TEAM_CD = ROLL.TEAM_CD
AND BASE.DEPT_CD = ROLL.DEPT_CD
— AND BASE.DEPT_CD = EA190086
GROUP BY SUBSTRING(BASE.DHSTC_CD,0,5),BASE.TEAM_CD,BASE.DEPT_CD,SUBSTRING(BASE.DHSTC_CD,5,8),BASE.STC_MONTH,BASE.H09_CNT,BASE.CVS_DEPT_PAX_CNT

)T
WHERE T.DH_YEAR = @YEAR AND T.DH_MONTH = @MONTH

/* 每个 TEAM 的统计 */

DECLARE @CVS_TEAM_STAT AS TABLE
(
TEAM_CD VARCHAR(20)
,GRP_CD VARCHAR(20)
,CHECK_RADIO numeric(19,2)
)
DECLARE @CVS_TEAM_BASE AS TABLE
(
DH_YEAR VARCHAR(20)
,STC_QUATR VARCHAR(20)
,STC_MONTH VARCHAR(20)
,DH_MONTH VARCHAR(20)
,TEAM_CD VARCHAR(20)
,DEPT_CD VARCHAR(20)
,CVS_GRP_PAX_CNT numeric(19,2)
,H09_CNT numeric(19,2)

)

INSERT INTO @CVS_TEAM_BASE
SELECT T1.DH_YEAR,T2.STC_QUATR,T1.DH_MONTH,T2.STC_MONTH,T2.TEAM_CD, – GRP_CD
,SUM(CVS_DEPT_PAX_CNT) CVS_GRP_PAX_CNT
,SUM(T2.H09_CNT) ACC_H09_CNT

FROM T_DM_DHSTC T1
,T_DM_DHSTC_DEPT T2
WHERE
–and TEAM_CD = EA190001
T1.DHSTC_CD = T2.DHSTC_CD
GROUP BY T1.DH_YEAR,T2.STC_QUATR,T1.DH_MONTH,T2.STC_MONTH
,T2.TEAM_CD

INSERT INTO @CVS_TEAM_STAT
SELECT T.TEAM_CD, – GRP_CD,T.CHECK_RADIO
FROM(
SELECT BASE.DH_YEAR,MAX(BASE.STC_QUATR) STC_QUATR,BASE.DH_MONTH,BASE.STC_MONTH,BASE.TEAM_CD,BASE.H09_CNT
,SUM(ROLL.H09_CNT) ACC
,BASE.CVS_GRP_PAX_CNT
,SUM(ROLL.H09_CNT)/BASE.CVS_GRP_PAX_CNT CVS_RAT
,CASE WHEN BASE.STC_MONTH = 01 THEN CAST(SUM(ROLL.H09_CNT)/(BASE.CVS_GRP_PAX_CNT*(33.0/100.0)) AS NUMERIC(19,2))
WHEN BASE.STC_MONTH = 02 THEN CAST(SUM(ROLL.H09_CNT)/(BASE.CVS_GRP_PAX_CNT*(66.0/100.0)) AS NUMERIC(19,2))
WHEN BASE.STC_MONTH = 03 THEN CAST(SUM(ROLL.H09_CNT)/(BASE.CVS_GRP_PAX_CNT*(100.0/100.0)) AS NUMERIC(19,2))
END CHECK_RADIO

FROM @CVS_TEAM_BASE BASE
,@CVS_TEAM_BASE ROLL
WHERE BASE.DH_YEAR = ROLL.DH_YEAR
–AND BASE.DH_MONTH = ROLL.DH_MONTH
AND BASE.STC_QUATR = ROLL.STC_QUATR

AND BASE.DH_MONTH = ROLL.DH_MONTH

AND BASE.TEAM_CD = ROLL.TEAM_CD

GROUP BY BASE.DH_YEAR,BASE.DH_MONTH,BASE.STC_MONTH,BASE.TEAM_CD,BASE.H09_CNT,BASE.CVS_GRP_PAX_CNT

)T
WHERE T.DH_YEAR = @YEAR AND T.DH_MONTH = @MONTH