Generating random numbers specific to different observations

[Chinmay Sharma]

I currently have data that looks like the following:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(IndividualNumber Flag)
1 1
1 0
1 0
1 0
2 1
2 0
2 0
2 0
3 1
3 0
3 0
4 1
4 0
4 0
4 0
end

wherein the IndividualNumber variable signifies ID of the individual. The above data is a subset of a bigger dataset, details of which are not really relevant for this question. Individuals are observed over several years. I wish to create a random variable drawn from a standard normal distribution specific to each individual over all years. In order to do this, I have first created a flag variable called Flag. I have used the putmata command to input these 2 vectors into Mata.

On Mata, I type the following code:

mata: D=J(rows(IndividualNumber), 1,0)

mata: for (Flag==1) {
matrix D[`i',1]=rnormal(1,1,0,1)

}

I am sure the code is horrendously wrong. I have first created a vector of zeroes, D. I wish to then replace elements of this vector by drawing from the standard normal distribution when the variable Flag takes on a value of 1. Any help is much apppreciated!

[Christophe Kolodziejczyk]

mata:

for (i=1;i<=rows(D);i++) {
      
       if (Flag[i,1]) D[i,1] = rnormal(1,1,0,1)

}

end

[Chinmay Sharma]

Thanks a lot! Really appreciate it.