Allow a case expression which takes two columns as an input about splink HOT 3 CLOSED

I find myself easily getting confused between (a) matching values derived from two columns (e.g. full_name) and (b) matching conditions based on two columns.

I'm trying to think of useful examples that aren't easily replicated by adding derived columns*, such as:

CASE WHEN l.first_name = r.first_name AND l.last_name = r.last_name THEN 2
WHEN l.first_name = r.last_name AND l.last_name = r.first_name THEN 1
ELSE 0 END

And this could still be referred to as gamma_name_swap or something even though name_swap isn't an actual column in the data.

*although I've just figured out how a derived column could solve this problem, and do it much better

from splink.

a derived column with both and then a jaccard sim on Q2grams would perhaps work here?
this is what im trying on swappy columns

from splink.

I was thinking of concatenating a sorted list of names (however you would do that in SQL) and matching on that:

John Adam Smith -> "AdamJohnSmith"
Adam John Smith -> "AdamJohnSmith"
Smith Adam Jonathan -> "AdamJonathanSmith"

from splink.