Question

我有一个数据集，要求我第一次找到注册的在线购物者购买的东西并对该购买应用5％的折扣。

数据集有28列但是出于这个问题的目的，我将把它浓缩为我认为相关的内容。

我需要创建一个新列，告诉我第一次有人购买了什么。我们可以假设在同一天进行的购买是相同的购买但属于不同的商品。

Obs     ID    Trans_Date   Order_Number   Value       Status
----------------------------------------------------------------
1874    866   30/07/2016   191            $4,217.90   Registered
1875    866   30/07/2016   191            $4,217.90   Registered
1876    866   31/07/2016   192            $2,422.75   Registered
1877    866   31/07/2016   192            $2,422.75   Registered
1878    .     31/07/2016   193            $4,162.66   Unregistered
1879    .     31/07/2016   193            $4,162.66   Unregistered
1880    344   31/07/2016   194            $4,405.51   Registered
1881    344   31/07/2016   194            $4,405.51   Registered
1882    .     31/07/2016   195            $2,114.76   Unregistered
1883    .     31/07/2016   195            $2,114.76   Unregistered
1884    250   31/07/2016   196            $3,310.72   Registered
1885    250   31/07/2016   196            $3,310.72   Registered
1886    .     31/07/2016   197            $4,633.48   Unregistered
1887    .     31/07/2016   197            $4,633.48   Unregistered
1888    .     31/07/2016   197            $4,633.48   Unregistered
1889    .     31/07/2016   197            $4,633.48   Unregistered
1890    .     31/07/2016   198            $6,224.43   Unregistered
1891    .     31/07/2016   198            $6,224.43   Unregistered
1892    .     31/07/2016   198            $6,224.43   Unregistered
1893    .     31/07/2016   198            $6,224.43   Unregistered

Answer 1

这是我的'第一个'数据集：

'obs'  ,   'id'  ,  'trans_date' ,  'order_number' ,  'value'      , 'status'
1874 ,   866 ,  30/07/2016 ,  191          ,  4217.90  , Registered
1875 ,   866 ,  30/07/2016 ,  191          ,  4217.90  , Registered
1876 ,   866 ,  31/07/2016 ,  192          ,  2422.75  , Registered
1877 ,   866 ,  31/07/2016 ,  192          ,  2422.75  , Registered
1878 ,   344 ,  30/07/2016 ,  193          ,  4162.66  , Unregistered
1879 ,   344 ,  30/07/2016 ,  193          ,  4162.66  , Unregistered
1880 ,   344 ,  31/07/2016 ,  194          ,  4405.51  , Registered
1881 ,   344 ,  31/07/2016 ,  194          ,  4405.51  , Registered
1882 ,   250 ,  30/07/2016 ,  195          ,  2114.76  , Unregistered
1883 ,   250 ,  30/07/2016 ,  195          ,  2114.76  , Unregistered
1884 ,   250 ,  31/07/2016 ,  196          ,  3310.72  , Registered
1885 ,   250 ,  31/07/2016 ,  196          ,  3310.72  , Registered
1886 ,   275 ,  30/07/2016 ,  197          ,  4633.48  , Unregistered
1887 ,   275 ,  30/07/2016 ,  197          ,  4633.48  , Unregistered
1888 ,   275 ,  30/07/2016 ,  197          ,  4633.48  , Unregistered
1889 ,   275 ,  30/07/2016 ,  197          ,  4633.48  , Unregistered
1890 ,   275 ,  31/07/2016 ,  198          ,  6224.43  , Unregistered
1891 ,   275 ,  31/07/2016 ,  198          ,  6224.43  , Unregistered
1892 ,   275 ,  31/07/2016 ,  198          ,  6224.43  , Unregistered
1893 ,   275 ,  31/07/2016 ,  198          ,  6224.43  , Unregistered

这里有一些proc sql：

proc sql noprint;
create table temp as
select *,min(trans_date) format=date9. as first
from first
group by id
order by order_number;

create table final as
select obs,id,trans_date,order_number,value,status,
   case when first = trans_date then 'FIRST'
        else 'NOT FIRST'
   end as flag
from temp;
quit;

查找变量的第一个实例

1 个答案: