Changing many variables with wildcard *

[Brian Phillips]

Hello all,

I am trying to change many variables in a time series data set, and would ideally like to do it in a manner such as "replace var* = var*[_n-1] if..." where the variables to be changed are var1 var2 var3 etc.

But Stata does not allow such use of the wildcard * in this context, and I understand why.

Is there a workaround? Perhaps using language such as foreach or forvalues, but I haven't found a way to do it with those commands yet.

For what it's worth, a manual version of what I am trying to do, that I would need to do for each of the hundreds of variables (intervention1, intervention2, etc.), looks like this:

by ccode: replace intervention6 = intervention6[_n-1] if intervention6[_n-1]!=0 & intervention6[_n-1]!=.
by ccode: replace intervention6 = 0 if startyear>intervention6
replace intervention6 = 1 if intervention6!=0

Where the unit of analysis is country-year, ccode=country code, and each variable indicates an intervention involving the country.

Thanks in advance.

[Clyde Schechter]

Perhaps using language such as foreach or forvalues, but I haven't found a way to do it with those commands yet.

You need to read those sections of the user manual more carefully . That precisely what they do.

foreach v of varlist var* {
    replace `v' = `v'[_n-1] if ...
}

is exactly what you are asking for.

[Clyde Schechter]

In fact, [P] on page 228 has examples like this, and the section is, in my view, very clearly written.

[Brian Phillips]

Thank you. I will look into that in more detail.