zoukankan      html  css  js  c++  java
  • 表的行列转置

    原始数据:

    select t1.CarModelPKID,t2.DefeatReason, count(*) as count from RT_CustAllocResult t
    left join vm_carinfo t1 on t1.FrameNo = t.FrameNo
    left join RT_CustFollowUp t2 on t2.RegNo = t.RegNo
    where t1.CarModelPKID is not null and t2.DefeatReason is not null group by t2.DefeatReason,CarModelPKID

    转置后的数据:

    select t.CarModelPKID,
    case DefeatReason when '00' then count end as colReason1,
    case DefeatReason when '01' then count  end as colReason2,
    case DefeatReason when '02' then count  end as colReason3,
    case DefeatReason when '03' then count  end as colReason4,
    case DefeatReason when '04' then count  end as colReason5,
    case DefeatReason when '05' then count  end as colReason6,
    case DefeatReason when '06' then count  end as colReason7,
    case DefeatReason when '07' then count  end as colReason8,
    case DefeatReason when '08' then count  end as colReason9
    
    from(select t1.CarModelPKID,t2.DefeatReason, count(*) as count from RT_CustAllocResult t
    left join vm_carinfo t1 on t1.FrameNo = t.FrameNo
    left join RT_CustFollowUp t2 on t2.RegNo = t.RegNo
    where t1.CarModelPKID is not null and t2.DefeatReason is not null group by t2.DefeatReason,CarModelPKID) t 

    数据统计:

    select t.CustomerLevel as Item,
    sum(case DefeatReason when '00' then count end) as colReason1,
    sum(case DefeatReason when '01' then count end) as colReason2,
    sum(case DefeatReason when '02' then count end) as colReason3,
    sum(case DefeatReason when '03' then count end) as colReason4,
    sum(case DefeatReason when '04' then count end) as colReason5,
    sum(case DefeatReason when '05' then count end) as colReason6,
    sum(case DefeatReason when '06' then count end) as colReason7,
    sum(case DefeatReason when '07' then count end) as colReason8,
    sum(case DefeatReason when '08' then count end) as colReason9
    
    from(select t.CustomerLevel,t2.DefeatReason, count(*) as count from RT_CustAllocResult t
    left join RT_CustFollowUp t2 on t2.RegNo = t.RegNo
    where t.CustomerLevel is not null and t2.DefeatReason is not null 
    group by t2.DefeatReason,t.CustomerLevel) t  group by t.CustomerLevel
  • 相关阅读:
    rasa learning to rank
    为什么选择rasa
    rasa
    tf.tile
    tf.scatter_nd
    nlp中的数据增强之 google-uda
    递归-分治-动态规划-贪心
    递归算法的美妙
    数据不平衡问题
    WD(西部数据)硬盘,“必须从您要解锁的硬盘对应的WD Drive Unlock CD 运行WD Drive Unlock应用程序”错误解决办法
  • 原文地址:https://www.cnblogs.com/ShaYeBlog/p/3594502.html
Copyright © 2011-2022 走看看