Collapse/Summarize Question

[Davia Downey]

I have US daily confirmed deaths and cases from the coronavirus pandemic that I need to collapse for use in another dataset. Johns Hopkins collected this data at the county level in each state, however I only need the statewide total. See snippet below:

* Example generated by -dataex-. For more info, type help dataex
clear
input float date2 str41 admin2 str24 provincestate long confirmed int deaths
21936 "Autauga" "Alabama" 0 0
21937 "Autauga" "Alabama" 0 0
21938 "Autauga" "Alabama" 0 0
21939 "Autauga" "Alabama" 0 0
21940 "Autauga" "Alabama" 0 0
21941 "Autauga" "Alabama" 0 0
21942 "Autauga" "Alabama" 0 0
21943 "Autauga" "Alabama" 0 0
21944 "Autauga" "Alabama" 0 0
21945 "Autauga" "Alabama" 0 0
21946 "Autauga" "Alabama" 0 0
21947 "Autauga" "Alabama" 0 0
21948 "Autauga" "Alabama" 0 0
21949 "Autauga" "Alabama" 0 0
21950 "Autauga" "Alabama" 0 0
21951 "Autauga" "Alabama" 0 0
21952 "Autauga" "Alabama" 0 0
21953 "Autauga" "Alabama" 0 0
21954 "Autauga" "Alabama" 0 0
21955 "Autauga" "Alabama" 0 0
21956 "Autauga" "Alabama" 0 0
21957 "Autauga" "Alabama" 0 0
21958 "Autauga" "Alabama" 0 0
21959 "Autauga" "Alabama" 0 0
21960 "Autauga" "Alabama" 0 0
end
format %td date2

I believe the correct collapse code would be something like collapse v1 v2...by(provincestate) HOWEVER, I need the actual counts and not a mean. Is there a command that will do this? Maybe Sum? Please advise.

[Ali Atia]

See -help collapse-. The default stat is mean, but you can also use sum:

collapse (sum) confirmed deaths,by(provincestate date2)

[Davia Downey]

What I need is to maintain the integrity of the daily numbers but for the entire state, not the counties. Will -collapse allow for that if I use -sum as part of the call?

[Ali Atia]

The command in #2 will add up all of the values for individual counties within a state on a given date such that the resulting dataset contains a single observation (summing up the individual county numbers) per state per date.