Tableau is probably not the tool for this job of trying to find missing or malformed raw data. You might be interested in trying out Project Maestro | Tableau Software though as it will help with some of these tasks.
Best of luck
@Patrick: I've thought about how to do this in Project Maestro and it's not a good fit because it doesn't have the iteration capability necessary for this kind of task. There's at least one brute force method I can think of that requires knowing how many products there could be. In any case Tableau isn't great for this either, honestly, but it's more theoretically feasible to do in Tableau.
@Christabelle: The second example doesn't make sense to me. How is one to know that Product 5 is the missing product from 20002/456 when it could just as easily be that Product 6 is missing from 20002/345? Also how many ID/ID2/Product combinations are there to evaluate? Hundreds, thousands, millions?
Thanks for replying, we don't know whether Product 5 or 6 is correct, we just need to find all the ones that don't match. The second table is an example where everything matches, and the first table are the erroneous examples which I need to find. I wish I had a better tool to find them, it's either Tableau or Excel for now. There are about 3000-4000, I think.
I might be missing something there, but couldn't this be done with a full outer Join where the keys are Id1, Id2, and Product? Or possibly even just a left join, if one of the tables represents the "master", and the other is the one we're looking for anomalies in? Then just look for NULL values on either side?