Hi,

I need to create the respondent-driven sampling variables identifying recruits from two variables:
recruiterID
and
participantID
.
caseID
recruiterID
participantID
1
project_id
Bn9Le8tQJt7uPT
2
Bn9Le8tQJt7uPT
VKERQF4fGz0LVY
3 Bn9Le8tQJt7uPT s0kRms5ndafnnfh
4
VKERQF4fGz0LVY
 R1E1MiDgYy8mcV
5 s0kRms5ndafnnfh f3kpXDlkxxgu7K
6
s0kRms5ndafnnfh
 yb960dHZgItYe4L

I need to add variables identifying caseID # 2 and #3 as being friend1 and friend2 of caseID #1, and caseID #5 and #6 as friend1 and friend 2 of caseID #3 (the other participants have missing values as they have not successfully recruited friends yet). For example:
caseID
recruiterID
participantID
 friend_code_1
friend_code_2
1
project_id
Bn9Le8tQJt7uPT
VKERQF4fGz0LVY
s0kRms5ndafnnfh
2
Bn9Le8tQJt7uPT
VKERQF4fGz0LVY
R1E1MiDgYy8mcV
3 Bn9Le8tQJt7uPT s0kRms5ndafnnfh
f3kpXDlkxxgu7K
 yb960dHZgItYe4L
4
VKERQF4fGz0LVY
 R1E1MiDgYy8mcV
5 s0kRms5ndafnnfh f3kpXDlkxxgu7K
6
s0kRms5ndafnnfh
 yb960dHZgItYe4L
Because the number of coupons differs according to some sociodemographic variables, I created
as much new friend_code_* variables as the maximum number of participants recruited by the same recruiter by the following syntax:

/*encoding str variables*/
encode recruiterID, generate(recruiterID_num)
encode participantID, generate(participantID_num)

/*counting and storing in r(max) the number of recruits to determine the number of friend_code_* variables to create*/
bysort recruiterID : gen nrecruits = _n
su nrecruits, meanonly

/*creating the friend_code_* variables*/
foreach i of num 1/`r(max)' {
by participantID_num, sort: gen friend_code_`i' = .
}


I need help figuring out how to copy the appropriate value of participantID in the friend_code_* variables (as required by the rds_network command). I just can't find the solution...

Many thanks,
Martin