Counting Efficiently

[Felix Stips]

Dear all,

I have a large dataset and want to count observations with particular characteristics. Precisely, I count the number of people in a municipality at that year from a certain nationality. My intuitive solution is to loop over these three dimensions and it works.

forvalues y = 2010(1)2015 {
    foreach k of local gemeinden {
        foreach c of local countries {
            count if origin == `c' & Gemeinde == "`k'" & Jahr == `y'
            replace network1Stock_gem = `r(N)' if origin == `c' & Gemeinde == "`k'" & Jahr == `y'
        }
    }
}

However, it is computationally expensive and Stata takes very long to perform the calculations. Therefore, I wanted to ask whether perhaps somebody has an idea to make the calculation more efficient?

Best,
Felix

[Nick Cox]

Seems to boil down to

bysort Jahr origin Gemeinde: gen wanted = _N

[Felix Stips]

lol, true. And it's much faster like that.. Thank you!