I've been struggling to find good info on how to do the following:
A(lice) and B(ob) each have a table of data (
b respectively). Each table has an
id column that is consistent between the parties.
id is a numeric with largeish but limited range. Otherwise the tables contain different data.
The tables have been deidentified apart from the
id. So it's impossible for A or B to derive what record they are looking at if they have
b without the
id column. This goes both for their own data and the other's data.
The parties wish to find the records that exist both in
b without revealing which ones matched.
Essentially, they need the result of
select * from a inner join b on a.id=b.id; with the id removed.
A & B may need to find a shared environment to do this, but the data must not leave A or B without it being anonymised.
I've got some amateur hour ideas, but I'm hoping experts here can help or point me in the right direction. I also hope that this is an appropriate question for this forum.