Generating a variable based on other variable

[Paris Rira]

Dear Profs and colleagues,

I am going to generate a variable that shows how many export varieties firms export.
Firm ID: NPC_FIC
Commodity Code: artigo
Type of trade (Export =2, Import=1): fluxo

* Example generated by -dataex-. For more info, type help dataex
clear
input double NPC_FIC str10 fluxo str24 artigo
10006 "Exporta��o" "01012990"
10006 "Exporta��o" "01012990"
10006 "Exporta��o" "01012990"
10006 "Exporta��o" "01012990"
10006 "Exporta��o" "01012990"
10006 "Exporta��o" "01012990"
10006 "Exporta��o" "01012990"
10006 "Exporta��o" "01012990"
500000001 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000002 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000033 ""           ""        
500000050 ""           ""        
500000050 ""           ""        
500000050 ""           ""        
500000050 ""           ""        
500000050 ""           ""        
500000050 ""           ""        
500000069 ""           ""        
500000069 ""           ""        
500000073 ""           ""        
500000073 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000083 ""           ""        
500000101 ""           ""        
500000101 ""           ""        
500000101 ""           ""        
500000101 ""           ""        
500000101 ""           ""        
500000101 ""           ""        
500000104 ""           ""        
500000106 ""           ""        
500000113 ""           ""        
500000113 ""           ""        
500000113 ""           ""        
500000113 ""           ""        
500000119 "Exporta��o" "07019090"
500000119 "Exporta��o" "07019090"
500000119 "Exporta��o" "07019090"
500000119 "Exporta��o" "07019090"
500000119 "Importa��o" "07019090"
500000119 "Importa��o" "07019090"
500000119 "Importa��o" "07019090"
500000119 "Importa��o" "07019090"
500000119 "Importa��o" "07019090"

Starting with :

by NPC_FIC fluxo (artigo), sort: gen n_ex = sum(artigo != artigo[_n-1])
by NPC_FIC fluxo (artigo): replace n_ex = n_ex[_N]

Could you please help me to finish that?

Cheers,
Paris

[Sebastian Schirner]

Hey Paris,

this could be version of code for you to consider. It provides the exact same results for your example.

unique

is a command available from ssc.

unique artigo if artigo!="", by(NPC_FIC fluxo) gen(n_ex)
bysort NPC_FIC fluxo (n_ex): replace n_ex = n_ex[1]

PS: Note, that when providing data there needs to be an

end

in the last line of your example.

Best
Sebastian

[Paris Rira]

It is fantastic, Thanks Sebastian, also for reminding me of END.
Though still --n_ex --does not distinguish the numbers of export. It shows export + import
After this stage, I need to create a variable to show the quantity of only export. Except --keep if fluxo==1 -- Do you know other way?

[Sebastian Schirner]

Hey Paris,

actually the variable captures the number seperate by import and export (since fluxo is included in by). In your current example one can not really see this. If you do the following after you input your example (random example) and run again you can see the difference.

replace artigo = "07019091" if _n ==69
unique artigo if artigo!="", by(NPC_FIC fluxo) gen(n_ex)
bysort NPC_FIC fluxo (n_ex): replace n_ex = n_ex[1]

I hope I am not getting something wrong.

Sebastian

[Paris Rira]

Hey Paris,

actually the variable captures the number seperate by import and export (since fluxo is included in by). In your current example one can not really see this.

You are right. It was labeled in that way indeed.
It has been dedicated number 1 for export and 2 for import.