creating a variable identyfying belonging to 2 other variables

[Guest]

Dear All,

I got stuck with my stata work. I have data on sectors (HSspecific) and countries (reporter). I wanna create a variable indicating that it is for a given country in a given sector, so that I have for example for Australia in sector 0 : AUS0. Here is a sample of my data:

* Example generated by -dataex-. To install: ssc install dataex
clear
input double tradevalue_ex str3 reporter long HSspecific
  315057867 "AUS" 0
   99401267 "AUS" 0
   30544335 "AUS" 0
   61603138 "AUS" 0
 2635123555 "AUS" 0
  132330437 "AUS" 0
 5267085798 "AUS" 0
  760438676 "AUS" 0
   49257876 "AUS" 0
   76051716 "AUS" 0
  457469316 "AUS" 0
   96004848 "AUS" 0
 1353869319 "AUS" 0
31251833517 "AUS" 0
  287704907 "AUS" 0
  160930750 "AUS" 0
  117388387 "AUS" 0
   87805534 "AUS" 0
 6158398175 "AUS" 0
   30460372 "AUS" 0
   52773899 "AUS" 0
   69735614 "AUS" 0
 3834472665 "AUS" 0
  831655308 "AUS" 0
  534397254 "AUS" 0
          0 "AUS" 0
15356473026 "AUS" 0
  248284111 "AUS" 0
   27206046 "AUS" 0
 1294341301 "AUS" 0
  828494978 "AUS" 0
  800932944 "AUS" 0
  197823700 "AUS" 0
 1031543916 "AUS" 0
17710695419 "AUS" 0
          0 "AUS" 0
 9814552469 "AUS" 0
 6114960461 "AUS" 0
  421347631 "AUS" 0
  277438187 "AUS" 0
   41767722 "AUS" 0
 5431902055 "AUS" 0
  129208194 "AUS" 0
 7069127457 "AUS" 0
   15777475 "AUS" 0
  313441983 "AUS" 0
   12640587 "AUS" 0
  707285627 "AUS" 0
 1420528572 "AUS" 0
  647406302 "AUS" 0
  752961734 "AUS" 0
          0 "AUS" 0
  704865737 "AUS" 0
   41337976 "AUS" 0
  557247237 "AUS" 0
   74196834 "AUS" 0
 1787465424 "AUS" 0
   59750768 "AUS" 0
 2382657934 "AUS" 0
   82395005 "AUS" 0
  375886871 "AUS" 0
   40981612 "AUS" 0
 1556555985 "AUS" 0
  529222435 "AUS" 0
  129210047 "AUS" 0
          0 "AUS" 0
 1798172822 "AUS" 0
 4925112283 "AUS" 0
  451818896 "AUS" 0
 4415983763 "AUS" 0
  666176798 "AUS" 0
  224357654 "AUS" 0
  303970724 "AUS" 0
   89367652 "AUS" 0
 5307657643 "AUS" 0
  100021130 "AUS" 0
  867382268 "AUS" 0
73804136765 "AUS" 0
 2980450888 "AUS" 0
 1029633820 "AUS" 0
 2447608055 "AUS" 0
 3687916642 "AUS" 0
 1126056378 "AUS" 0
  587539406 "AUS" 0
   23610568 "AUS" 0
   42009484 "AUS" 0
  366882895 "AUS" 0
 1429329689 "AUS" 0
 7356810011 "AUS" 0
10184268115 "AUS" 0
  519814575 "AUS" 0
  318655944 "AUS" 0
   33176840 "AUS" 0
  847104126 "AUS" 0
   11142131 "AUS" 0
  579277761 "AUS" 0
 2778133661 "AUS" 0
  134253261 "AUS" 0
  203037891 "AUS" 0
  135200284 "AUS" 0
end

I was just trying to multiply the variables like:

gen country_sector=HSspecific*reporter

but it didn't yield the expected results. Then I tried to destring, encode etc, but wasn't successful. I've just run out of ideas and hope You could help me

[Andrew Musau]

egen wanted= group(reporter HSspecific), label
tab wanted

[Guest]

Thank you for your answer. I tried it, but unfortunately I got an error message: too many values for label define. I have panel data for 26 countries from 1990 till 2018 on export and import for almost 900 sectors.I needed to calculate aggregate trade weights for every country in a given sector, where weight is a country share in trade in a given sector. I calculated the sum of export for a given sector by applying:

bysort HSspecific: egen sum_export=total(tradevalue_ex)

The sum for a given country by:

bysort HSspecific reporter: egen sum_export_reporter=total(tradevalue_ex)

Then I generated the variable weight:

gen weight=sum_export_reporter/sum_export

My idea was to have a column with the variables identyfying the country and sector at once, like AUS0 e.g.

[Andrew Musau]

You can have a string-identifier then:

gen wanted= reporter +"-"+HSspecific
tab wanted in 1/200

[Guest]

Stata showed me an error: type mismatch. When I clicked in the error I saw that I wanted to combine a string and numeric subexpression. I didn't want to try your patience since you helped me a lot before. After some struggles I got the command:

egen countrysector = concat(reporter HSspecific)

It worked! But thank you for your answer anyway

[Andrew Musau]

Sorry, should be

gen wanted= reporter +"-"+string(HSspecific)

It is no bother. Always express any difficulties that you may have.