Finding the average age of death in a database in which the same person appears multiple times

[Emanuele Zanon]

Hello, I have a most likely banal question. I have a database of about 400 000 observations which contains the following information:

ID	age_of_death
1	50
1	50
1	50
2	.
2	.
3	60

A missing value means that the person is alive, so their age should not enter my calculation at all. I want to calulate the average age at death, but I don't know how to consider only one value for each ID and ignore the missing values. I assumed I needed to use the tag command, but I can't figure out exactly how.
For clarity, if the above data was my data, the value I need would be: average_age_at_death = (50 + 60)/2

[Nick Cox]

tag() is an egen function, not a command. That aside, it can be part of a solution, and the other part is that summarize ignores missings any way.

clear 
input ID    age_of_death
1    50
1    50
1    50
2    .
2    .
3    60
end 

egen tag = tag(ID)
su age_of_death if tag

[Maarten Buis]

Have all persons in your dataset died, or are some (hopefully many) of your respondents still alive when the data collection stopped, and thus their age of death unknown?