Help egen anyvalue with another variable instead of a integer list.

[Marco Tomasino]

In a datasets with two variables, Var1 and Var2, how can I drop the records that have a Var1 value different to any Var2?

I thought something like a egen with anyvalue, but the ¨integer list¨ is composed by all the values of Var2.

Thanks

[Ken Chui]

Welcome to Statalist. Could you elaborate this a bit more, and preferably with an example?

What does it mean by "any Var2"? Do you mean:

1) Var 2 can have values, for example, 1, 2, 3, 4, 5; and if Var 1 is not any of these these 5 values, drop the case.

or

2) Drop the case if Var 1 is different that Var 2?

Please take a moment to read the FAQ (http://www.statalist.org/forums/help) and check out how to use dataex to provide an example. That'd speed up the exchanges.

[Marco Tomasino]

Thanks Ken

I think the case would be the the first one 1): Var 2 can have values, for example, 1, 2, 3, 4, 5; and if Var 1 is not any of these these 5 values, drop the case.

clear
input float ID long stub1
1 .
2 .
3 .
4 .
5 .
6 .
7 2
8 5
9 7
10 .
end

In this example, I'd like to keep the records with ID 2 5 and 7 and drop the rest.

[Clyde Schechter]

frame put stub1, into(to_match)
frame to_match {
    drop if missing(stub1)
    duplicates drop
    isid stub1, sort
}

frlink m:1 ID, frame(to_match stub1)
keep if !missing(to_match)
drop to_match
frame drop to_match

[Marco Tomasino]

Thank you very much Clyde! That was very helpful.

And how can I make Stata keep not only ID 2, 5 and 7, but also 8 and 9 of the example?
That would be, to drop the records that both have
1) ID not included in any of the stub1 values
2) stub1 ==.

Thank you!

[Clyde Schechter]

Change -keep if !missing(to_match)- to -keep if !missing(to_match) | !missing(stub1)-.