1

我有以下数据集：

需要添加两个新列 - 第一个从每个客户的第 2 行中减去第 1 行，这样我们就可以获得客户续订会员资格的“天数”数 - 第二个计算客户续订会员资格的次数这只是从 0 开始的计数。

Row   - Customer    - Renew Date   - Type of Renewal  - Days_Since  -Prev_Renewal
 1       - A        - June 10, 2010        - X                        
 2       - A        - May 01, 2011         - Y  
 3       - B        - Jan 05, 2010         - Y  
 4       - B        - Dec 10, 2010         - Z   
 5       - B        - Dec 10, 2011         - X

这是我现在正在使用的代码。有没有办法将这两组查询组合成一个？

data have;
informat renew_date ANYDTDTE.;
format renew_date DATE9.;
infile datalines dlm='-';
input Row Customer $ Renew_Date  Renewal_Type $;
datalines;
 1       - A        - June 10, 2010        - X                        
 2       - A        - May 01, 2011         - Y  
 3       - B        - Jan 05, 2010         - Y  
 4       - B        - Dec 10, 2010         - Z   
 5       - B        - Dec 10, 2011         - X    
 ;;;;
 run;

 data want;
 set have;
 by customer;
 retain prev_days;  *retain the value of prev_days from one row to the next;
 if first.customer 
 then 
    days_since=0; 
    *initialize days_since to zero for each customer's first record;
 else days_since=renew_date-prev_days; *otherwise set it to the difference;
 output;    *output the current record;
 prev_days=renew_date; 
 *now change prev_days to the renewal date so the next record has it;
 run; 

data want1;  
 set have;  
by customer;   
retain prev_renewal;   
if first.customer then prev_renewal=0;   
else prev_renewal=prev_renewal+1;   
output;   
run;

谢谢

4

1 回答 1

0

这不是 SQL，而是数据步骤/基本 SAS 代码——SAS Institute 的专有 (4GL) 语言，它实际上早于 SQL。

关于您的程序 - 值得指出的是，在使用 BY 语句之前，必须对数据进行排序或索引（在这种情况下由客户）。在这种情况下，您的数据线顺序正确。

这是您需要的组合代码：

data want (drop=prev_days);
   set have;
   by customer;
   retain prev_days; 
   if first.customer then do;
      days_since=0; 
      prev_renewal=0;
   end;
   else do;
      days_since=renew_date-prev_days;
      prev_renewal+1; /* IMPLIED retain - special syntax */
   end;
   output;
   prev_days=renew_date;
run;
于 2014-02-22T13:27:24.753 回答