List unique obs from appended dataset

[abhishek bhati]

clear
input long ID str100 org_name str16 datasource
1234456 "Helping Hand" "Survey1"
1234456 "Helping Hand" "Survey2"
9089081 "Save Children" "Survey1"
9089081 "Save Children" "Survey2"
9089081 "Save Children" "Survey3"
4532901 "Planet Earth" "Survey1"
end

I am interested in creating a list of organizations that have participated in all three surveys survey1, survey2 & survey3 (variable - data source). So basically, knowing the # of organizations that participated in three surveys, or just two or one.

In the end, a new dataset from the master dataset of organizations based on data source based on the organizations which took all three surveys, or two, etc.

Thanks in advance!

[Andrew Musau]

Something like:

isid org_name datasource
preserve
contract ID
*ALL 3
list if _freq==3
*TWO
list if _freq==2
restore

[abhishek bhati]

Its giving me an error message that org_name is not uniquely defined which it is not. The unique identifier is ID. I was wondering rather than listing the ID, could I create a new dataset of organizations that participated in all three surveys. Output like
9089081 "Save Children" "Survey1"
9089081 "Save Children" "Survey2"
9089081 "Save Children" "Survey3"

[Andrew Musau]

If the following does not do it, you need to give a data example using dataex as your description will appear not to match the actual dataset. See FAQ #12 for more details.

preserve
keep ID org_name datasource
duplicates drop *, force
bys ID: drop if _N<3
l, sepby(ID)
restore

[abhishek bhati]

Thanks, Andrew. The code is working and lists the observation. However, is it possible to drop rest of the observations rather than listing them. I am interested in creating a new dataset.

[Andrew Musau]

forval i=1/3{
    bys ID: egen in`i'= max(datasource=="Survey`i'")
}
preserve
keep if in1&in2&in3
save all3, replace
restore
use all3, clear