When I'm querying data sets, and I'm trying to get a better understanding of a population, I'll often filter my results down to one individual as a starting point. Select all records in multiple joined tables where id number equals 'blahblah...this person's id number'
I come to know so much about this one individual. I can then make some pretty good guesses, some pretty good assumptions about what the other members of the population might be like. It's a good test for when you're trying to pull a list of individuals who fit some set of criteria. I get these sets of criteria from some manager or director or 'analyst' who thinks it is important. Maybe it is or maybe it isn't. These sets of criteria are often vague, not very well uninformed. I don't fault the requestor for that, though. They are neck deep in their business, just like I am single-nostril deep in mine. They don't have time to know what I know, just like I don't have time to know what they know.
Anyway, when I think I've got enough tables joined together to tell some 'story' to some analyst -- what events these subjects have been to, what kinds of memberships they have, how much they paid on their last invoice, what they subscribe to...blah blah, I'll take the filtering off of that one individual and pull in all the individuals who fit into this criteria and see what I get. I see commonalities. I see how all their data is joined.
An important thing that happens at this point is: I start to see individuals who definitely do not fit the criteria. I find ways to filter them out. You have to have a record in this one last table, or you don't make it into the set.
I don't really care at all about the one person I started with. I'm just trying to gain a bigger understanding of everybody's records in that whole population. Once I've pulled the full list of desired individuals, I don't even remember what that first person was all about. As a matter of fact, after reflection during the process, that first individual might not even make it into the final set! I might find one little detail about them that...causes them to be...an exclusion.
All of this doesn't have anything to do with anything -for me- except pulling data from relational databases at work. It doesn't seem like a good way to be -in real life- going around 'fact gathering' on people and eliminating them based on one or two data fields.