Drop duplicate pairs (X,Y; Y,X)

[Harald Leber]

Unfortunately, I was not able to find out how to drop a pair of duplicates. Let's say we have variables ID, X and Y and the data looks as follows:

ID	X	Y
1	1	2
2	2	1
3	3	5
4	4	6

I would like to identify ID1 and ID2 as a pair and drop ID2. However, this does not work with "duplicates". Do you have an efficient solution for this problem?

Thank you!

[Rich Goldstein]

you could use the user-written "rowsort" (SJ; use search to find and install) and then use duplicates (and then drop the new variables); see;

help rowsort

[Nick Cox]

https://www.stata-journal.com/articl...article=dm0043 explains.

(Searching for mentions of dm0043 here will give you examples.)

[Harald Leber]

Works like a charm, thanks!

[David Benson]

Harald, sounds like you got what you needed.

For those coming afterwards, I will also point to two other posts here on Statalist that answer similar questions regarding creating a unique group or dyad identifier (they all use the same solution given in the Stata Journal article): See here and here