Turning three variables into one based on year. Code included

[Tyler Spencer]

* Example generated by -dataex-. For more info, type help dataex
clear
input double hhid int intyear double(dst2012 dst2010 dst2008)
1010140020171 2008 .71 .71 4.124
1010140020171 2010 .71 .71 4.124
1010140020171 2012 .71 .71 4.124
1010140020284 2008 .71 .71 4.124
1010140020284 2010 .71 .71 4.124
end

Hi there,

I working to build a panel dataset and I've run into a bit of trouble figuring out how to create a single distance variable from three separate distance variables based on year. I need a variable that will pull in the correct distance based on the year variable. E.g., in the first observation, I'd like the hhid for 2008 to correspond to a single variable that would pull in the dst2008 first observation, while the next hhid observation would correspond to dst2010. Any help would be greatly appreciated.

Kind regards,

[Clyde Schechter]

It is not clear to me whether you want to retain the three dst* variables and add one more that has the year-intyear correspondence:

reshape long dst, i(hhid intyear) j(year)
by hhid intyear: egen wanted = max(cond (year == intyear, dst, .))
reshape wide

or if you want to just end up with a single dst variable extracted from the one corresponding to the value of intyear, losing the original dst* variables:

reshape long dst, i(hhid intyear) j(year)
keep if year == intyear
drop year