求一小段stata代码，如图所示

如图，A的keyword，最早出现“1”的年份是2001，所以A的want=2001

同时，B的keyword，最早出现“1”的年份是2003，所以B的want=2003

示例文件见附件
stata该如何写代码？




示例数据如下：

1. * Example generated by -dataex-. To install: ssc install dataex
   
2. clear
   
3. input str1 ID int year str1 keyword float want
   
4. "A" 2000 "0" 2001
   
5. "A" 2001 "1" 2001
   
6. "A" 2002 "0" 2001
   
7. "A" 2003 "1" 2001
   
8. "A" 2004 "1" 2001
   
9. "A" 2005 "1" 2001
   
10. "B" 2000 "0" 2003
    
11. "B" 2001 "0" 2003
    
12. "B" 2002 "0" 2003
    
13. "B" 2003 "1" 2003
    
14. "B" 2004 "0" 2003
    
15. "B" 2005 "1" 2003
    
16. end

复制代码

求出现该ID下keyword最早出现的年份.dta (3.14 KB)

2019-11-16 09:37:37 上传

bys ID (year): g w=(sum(real(keyword))==1)*year*(sum(real(keyword[_n-1]))==0)
by ID: egen wa=max(w)

destring keyword, replace
bys ID:gen s = sum(keyword)
gen wanti=year if keyword == 1 & keyword[_n-1]!=1 & s==1

sungmoo 发表于 2019-11-16 12:59
bys ID (year): g w=(sum(real(keyword))==1)*year*(sum(real(keyword[_n-1]))==0)
by ID: egen wa=max(w)

非常感谢，不知道bys ID 和by ID，有什么区别？

diannaoasd 发表于 2019-11-16 12:19
destring keyword, replace
bys ID:gen s = sum(keyword)
gen wanti=year if keyword == 1 & keyword[_n- ...

或试试

1. destring keyword, replace
   
2. gen tem = year*keyword if keyword == 1
   
3. bys ID: egen wanted = min(tem)

复制代码

黃河泉 发表于 2019-11-16 15:54
或试试

非常感谢您~前几天没有登录论坛，所以回复迟了，谢谢~~

diannaoasd 发表于 2019-11-19 19:40
非常感谢您~前几天没有登录论坛，所以回复迟了，谢谢~~

No problem at all.