Jeremy: On Tue, May 21, 2019 at 11:58 PM Jeremy Finzel <finzelj@xxxxxxxxx> wrote: > Then take Francisco's suggestion, only use an md5 of the organization fields to create yourself a unique identifier. Then you can use ctid (unique internal identifier for each row) to join back. You use SQL like this: Sadly my suggestion only works if you can ( manually ) assign an organization line to a people line, md5, field concatenation, everything else is just optimization. >From what the OP has already told you have a heap of people, a heap of organizations and a magic device to assign one to the others, the org_id assignment ( using an unknown algorithm, we do not know if he wants sequences, texts or cat gifs as IDs ) is easy, the pairing part is unsolvable with the data we have. At this moment I think the only useful link for this is http://xyproblem.info/ ( for the OP, not for U ). Francisco Olarte.
