Create variable for group number in a country per year

[Santiago Molano]

Hi! I have three variables, namely group, country and year. Each country has x groups, which can change depending on what year one observes. What I am trying to do then is creating a variable that tells me the number of groups in a given country in a given year. The data of the three variables looks like this:

* Example generated by -dataex-. For more info, type help dataex
clear
input str84 group str37 country float year
"Balanta" "Guinea-Bissau" 1990
"Balanta" "Guinea-Bissau" 1991
"Balanta" "Guinea-Bissau" 1992
"Balanta" "Guinea-Bissau" 1993
"Balanta" "Guinea-Bissau" 1994
"Balanta" "Guinea-Bissau" 1995
"Balanta" "Guinea-Bissau" 1996
"Balanta" "Guinea-Bissau" 1997
"Balanta" "Guinea-Bissau" 1998
"Balanta" "Guinea-Bissau" 1999
end

Thanks in advance!

[Carlo Lazzaro]

Santiago:
do you mean somethìng along the following lines?

. bysort country year: g wanted=_n

. list

     +-----------------------------------------+
     |   group         country   year   wanted |
     |-----------------------------------------|
  1. | Balanta   Guinea-Bissau   1990        1 |
  2. | Balanta   Guinea-Bissau   1991        1 |
  3. | Balanta   Guinea-Bissau   1992        1 |
  4. | Balanta   Guinea-Bissau   1993        1 |
  5. | Balanta   Guinea-Bissau   1994        1 |
     |-----------------------------------------|
  6. | Balanta   Guinea-Bissau   1995        1 |
  7. | Balanta   Guinea-Bissau   1996        1 |
  8. | Balanta   Guinea-Bissau   1997        1 |
  9. | Balanta   Guinea-Bissau   1998        1 |
 10. | Balanta   Guinea-Bissau   1999        1 |
     +-----------------------------------------+

.

[Nick Cox]

This is a variation on your previous question https://www.statalist.org/forums/for...oups-per-state

See https://www.stata-journal.com/articl...article=dm0042 for a review of this question. dm0042 is thus revealed as an otherwise unpredictable search term for this forum and otherwise, as is the command name distinct.

egen tag = tag(group country year) 
egen ngroups = total(tag), by(country year)

is a way to make progress here, and more discussion can be found in the cited paper.

[Santiago Molano]

Santiago:
do you mean somethìng along the following lines?

. bysort country year: g wanted=_n

. list

+-----------------------------------------+
| group country year wanted |
|-----------------------------------------|
1. | Balanta Guinea-Bissau 1990 1 |
2. | Balanta Guinea-Bissau 1991 1 |
3. | Balanta Guinea-Bissau 1992 1 |
4. | Balanta Guinea-Bissau 1993 1 |
5. | Balanta Guinea-Bissau 1994 1 |
|-----------------------------------------|
6. | Balanta Guinea-Bissau 1995 1 |
7. | Balanta Guinea-Bissau 1996 1 |
8. | Balanta Guinea-Bissau 1997 1 |
9. | Balanta Guinea-Bissau 1998 1 |
10. | Balanta Guinea-Bissau 1999 1 |
+-----------------------------------------+

.

Yes, exactly!

[Santiago Molano]

This is a variation on your previous question https://www.statalist.org/forums/for...oups-per-state

See https://www.stata-journal.com/articl...article=dm0042 for a review of this question. dm0042 is thus revealed as an otherwise unpredictable search term for this forum and otherwise, as is the command name distinct.

egen tag = tag(group country year)
egen ngroups = total(tag), by(country year)

is a way to make progress here, and more discussion can be found in the cited paper.

Worked perfectly, thank you for the help

[Nick Cox]

#2 and #3 aren't the same answer in principle. They will give the same answer for the data example.