Merging on country pairs regardless of order

[Anna Star]

Hello,

I need to merge two datasets, each of which have two variables that together identify a country pair: e.g. country1=="USA"and country2=="CAN". My issue is that I need them merged regardless of the whether they appear like this or in reverse, i.e. country1=="CAN" and country2=="USA".

I played around creating a single identifier variable, but I couldn't figure it out.

Any ideas will be greatly appreciated!

[Nick Cox]

Discussed at https://www.stata-journal.com/sjpdf....iclenum=dm0043

dm0043 is thus a search term for several related threads.

[Anna Star]

Thank you! With some modification of the conditional statement, it worked - the key was to know that < and > work alphabetically with strings.