Interviewers would be better off focusing on data fundamentals: data modeling, validation, schema design, indexing, sharding/distribution strategies. When you screw this stuff up, it can be very, very hard to fix. The "problems" cascade and build up. Garbage in, garbage out. I've seen databases, in production, with no primary keys or indexes. It is absolutely bonkers.