aving with filename defined by value in a cell

[Ciaran OFlynn]

Saving with filename defined by value in a cell

Hi,

I have a dataset

* Example generated by -dataex-. For more info, type help dataex

clear

input str18 bvd_id_number str2 countrycode

"PT503338435" "PT"

"MA37163-81" "MA"

"JP1011001052846" "JP"

"FI10958757" "FI"

"US115044924GN" "US"

"US112599986GN" "US"

"IT03704400237" "IT"

"IT01868030543" "IT"

"HU14010040" "HU"

"PT501850066" "PT"

end

I'm using a loop to split it into country sets (master dataset is approx 400gb .csv).




the stata code I'm using is

egen group = group(countrycode)

su group, meanonly

foreach i of num 1/`r(max)' {
 preserve   keep if group == \`i'  save \`i'.dta  
restore

}

This works fine to split the file into individual countries, but it saves the file as file1.dta, file2.dta, etc. This wouuld be ok, except I will need to append multiple files together to make a whole country dataset, and file1.dta is not always consistently AD (Andorra), it might be AE(United Arab Emirates), for example.

How can I tell Stata to use any value from the countrycode variable that it has in memory to use as the filename?




Thanks

[Clyde Schechter]

There is no need for the variable group here. If you need it for some other purpose, fine, but don't use it as the iterator in your loop.

levelsof countrycode, local(countries)
foreach c of local countries {
    preserve
    keep if countrycode == "`c'"
    save `c', replace
    restore
}

Note: Assumes that the country codes do not contain any embedded blanks. If they do, then -save `"`c'.dta"', replace-. Also assumes the country codes do not contain any embedded quotation marks. If they do, -keep if countrycode == `"`c'"'-.

[Ciaran OFlynn]

Super - thank you!