The set operators are:
These are very useful in many kinds of queries. I use them extensively in Detecting Changed Data and in Data Quality queries.
In all the implementations I know, the set operators are particularly efficient, being consistently faster than joins (which perform a similar, but importantly different, function).
Hive does provide a "union" operator, but it does not provide the other set level operators:
This omission makes it significantly more difficult to write certain queries, and therefore increases development costs.