What to do with problems in a large inventory?

What I have been doing lately with my project is doing a pin by pin inventory of my collection. I put this off as long as I could, but I finally decided it had to be done, even though I don't actually have possession of my entire collection at the moment. There have been several instances in the past where I realized something was wrong, but just told myself -- I'll figure it out later. I guess now is later. I've already gone through about 7 or 8 smallish genera and have already found a problem. I have two ants on two different pins with the same ID number. In addition, I have a record in my database for the same species but with a different ID number, which I cannot find a physical specimen for. So, you would think, obviously one of those ants is actually the other ID number, but which one? They were both collected with the same collection method, but from different transects. They are the same species, at least, but what is the right thing to do here? Not count either one in my statistics? Just randomly choose one to be the other ID number and assume that it doesn't matter because they are the same species? Delete the record for the specimen I can't find? In an inventory with such a large number of specimens, maybe it doesn't matter what I do with one single specimen. But of course I believe that it does matter -- and who knows how many other problems I will eventually find. In this case it is a fairly common species, but what if it happens with a species that I only have one or two specimens for? Whatever I do might drastically change my statistics? I have no idea at this point, I'm just wondering if anyone out there has encountered similar issues before.