I mean, fwiw sometimes "obscure data structures" really do cleanly, efficiently, and performatively solve a problem that would have been a nightmare to address with simpler, more well known structures. That said, for most of us, it's rare enough to be memorable - but it does happen.
I think you're right that requiring such knowledge off the top of one's head and in the context of a one-hour interview is almost always unreasonable, though.
What I'd really want to know is how well a person can do the research to arrive at performant solutions involving, as you say, obscure data structures. In an interview context though, I think you basically just have to ask them - would be nice if you could get a sense of it through doing, though.
I think you're right that requiring such knowledge off the top of one's head and in the context of a one-hour interview is almost always unreasonable, though.
What I'd really want to know is how well a person can do the research to arrive at performant solutions involving, as you say, obscure data structures. In an interview context though, I think you basically just have to ask them - would be nice if you could get a sense of it through doing, though.