Repeat the value after the first appearance

[Andy Gonzalez]

Hello! I hope you can help me with the next task

Here is an extract of my data set

The variable trans adopts the value of 0 or 1. After the first appearance of “1” in the variable trans for a product code for a country, the value of trans should be 1 for the rest of the years for that product code for that country. For example, if the first time the trans value is 1 for product code 721070 of country 410 is in 1997, then from 1998 to 2021, the value of trans should be replaced by 1. My data set goes from 1995 to 2021, I have 20 countries and thousands of different product codes.

Year	Country	ProductCode	trans
1996	410	721050	1
1996	410	721060	0
1996	410	721070	0
1997	410	721050	1
1997	410	721060	1
1997	410	721070	1
1998	410	721050	0
1998	410	721060	0
1998	410	721070	0
1996	230	721050	0
1996	230	721060	0
1996	230	721070	0
1997	230	721050	1
1997	230	721060	1
1997	230	721070	1
1998	230	721050	0
1998	230	721060	0
1998	230	721070	0

Thank you in advance [COLOR=inherit !important]


[/COLOR]

[Nick Cox]

Please use dataex as here and as requested in FAQ Advice #12.

I wouldn't overwrite an existing variable unless you are completely sure you'll never want it again and that your replacement code is exactly right. Even then someone else may want to see the original data.

This may help otherwise.

* Example generated by -dataex-. For more info, type help dataex
clear
input int(year country) long productcode byte trans
1996 410 721050 1
1996 410 721060 0
1996 410 721070 0
1997 410 721050 1
1997 410 721060 1
1997 410 721070 1
1998 410 721050 0
1998 410 721060 0
1998 410 721070 0
1996 230 721050 0
1996 230 721060 0
1996 230 721070 0
1997 230 721050 1
1997 230 721060 1
1997 230 721070 1
1998 230 721050 0
1998 230 721060 0
1998 230 721070 0
end

gen wanted = trans 
bysort country productcode (year) : replace wanted = 1 if wanted[_n-1] == 1

[Andy Gonzalez]

Many many thanks Nick! It is working