多个时间间隔内分组累计函数的计算

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

我有如下的一个数据（以下是一部分样本，并且同样的数据集有多个），各个变量含义依次是股票代码、交易日期、某个收益值、基准日期（记为第0天）、基准日前两天（记第-2天）、基准日后两天（记第2天）、后边的依次是后4、6、8、10天（4、6、8、10），现在需要计算不同股票在以下时间区间内【-2，2】、【0，2】、【0，4】，【0，6】，【0，8】，【0，10】跟ar有关的一个函数值，请教该如何写程序？或者请高手提示下该用什么？循环和宏吗？
stck Trddt ar date pre2_date aft2_date aft4_date aft6_date aft8_date aft10_date
000009 2008-03-06 0.02442681 2008-03-08 2008-03-06 2008-03-10 2008-03-12 2008-03-14 2008-03-16 2008-03-18
000009 2008-03-07 -0.020126482 2008-03-08 2008-03-06 2008-03-10 2008-03-12 2008-03-14 2008-03-16 2008-03-18
000009 2008-03-10 -0.021732947 2008-03-08 2008-03-06 2008-03-10 2008-03-12 2008-03-14 2008-03-16 2008-03-18
000026 2008-04-07 -0.035743757 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-08 -0.021548374 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-09 0.021731954 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-10 -0.031038653 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-11 0.000708965 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-14 -0.027964365 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-15 -0.048091734 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-16 0.025881512 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-17 -0.007311003 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
000026 2008-04-18 -0.022120984 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：时间间隔 date 股票代码 ATE 基准日 函数 分组 时间间隔

1. data a;
   
2.         input
   
3.         stck:$6. Trddt:yymmdd10. ar date:yymmdd10. pre2_date:yymmdd10. aft2_date:yymmdd10. aft4_date:yymmdd10. aft6_date:yymmdd10. aft8_date:yymmdd10. aft10_date:yymmdd10.;
   
4.         format Trddt date pre2_date aft2_date aft4_date aft6_date aft8_date aft10_date yymmdd10.;
   
5.         datalines;
   
6. 000009 2008-03-06 1 2008-03-08 2008-03-06 2008-03-10 2008-03-12 2008-03-14 2008-03-16 2008-03-18
   
7. 000009 2008-03-07 2 2008-03-08 2008-03-06 2008-03-10 2008-03-12 2008-03-14 2008-03-16 2008-03-18
   
8. 000009 2008-03-10 3 2008-03-08 2008-03-06 2008-03-10 2008-03-12 2008-03-14 2008-03-16 2008-03-18
   
9. 000026 2008-04-07 4 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
   
10. 000026 2008-04-08 5 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
11. 000026 2008-04-09 6 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
12. 000026 2008-04-10 7 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
13. 000026 2008-04-11 8 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
14. 000026 2008-04-14 9 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
15. 000026 2008-04-15 10 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
16. 000026 2008-04-16 11 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
17. 000026 2008-04-17 12 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
18. 000026 2008-04-18 13 2008-04-09 2008-04-07 2008-04-11 2008-04-13 2008-04-15 2008-04-17 2008-04-19
    
19. ;
    
20. run;
    
21. 
    
22. %MACRO Cal(min,max);
    
23.         %if &min<0 %then %let var_min=pre%eval(-&min)_date;
    
24.         %else %if &min=0 %then %let var_min=date;
    
25.         %else %let var_min=aft&min._date;
    
26. 
    
27.         %if &max<0 %then %let var_max=pre%eval(-&max)_date;
    
28.         %else %if &max=0 %then %let var_max=date;
    
29.         %else %let var_max=aft&max._date;
    
30. 
    
31.         %put &var_min &var_max;
    
32. 
    
33.         ,sum(case when Trddt>=&var_min and Trddt<=&var_max then ar else 0 end) as %scan(&var_min,1,_)_%scan(&var_max,1,_)
    
34. %MEND Cal;
    
35. 
    
36. 
    
37. 
    
38. proc sql;
    
39.         create table b as
    
40.                 select
    
41.                         stck
    
42.                         ,date
    
43.                         %Cal(-2,2)
    
44.                         %Cal(0,2)
    
45.                         %Cal(0,4)
    
46.                         %Cal(0,6)
    
47.                         %Cal(0,8)
    
48.                         %Cal(0,10)
    
49.                 from a
    
50.                 group by
    
51.                         stck
    
52.                         ,date
    
53.         ;
    
54. quit;

复制代码

抛个砖，用sum做的测试，其他聚合函数的话要稍作改动

用循环、IF 语句能不能做啊？

没用过宏，学的话时间有点紧。。。

学习了，高效的编程

这个宏的确赞，不过核心思想是：挑选符合条件的观测。
我做的话，会些类似如下的code：
data b;set a;
if abs(intck('day',trddt,date))<=2 then _minus2_to_2=1;else _minus2_to_2=0;
if intck('day',trddt,date)>=0 and intck('day',trddt,date)<=2 then _0_to_2=1;else _0_to_2=0;
if intck('day',trddt,date)>=0 and intck('day',trddt,date)<=4 then _0_to_4=1;else _0_to_4=0;
if intck('day',trddt,date)>=0 and intck('day',trddt,date)<=6 then _0_to_6=1;else _0_to_6=0;
if intck('day',trddt,date)>=0 and intck('day',trddt,date)<=8 then _0_to_8=1;else _0_to_8=0;
if intck('day',trddt,date)>=0 and intck('day',trddt,date)<=10 then _0_to_10=1;else _0_to_10=0;
run;
然后再用sql汇总。