We review the notion of hypertree width, a measure of the degree of cyclicity of hypergraphs that is useful for identifying and solving efficiently easy instances of hard problems, by exploiting their structural properties. Indeed, a number of relevant problems from different areas, such as database theory, artificial intelligence, and game theory, are tractable when their underlying hypergraphs have small (i.e., bounded by some fixed constant) hypertree width. In particular, we describe how this notion may be used for identifying tractable classes of database queries and answering such queries in an efficient way.