Spark is distributed in two ways: either as precompiled binaries or as a source code that gives you the flexibility to choose, for example, whether you need support for Hive or not. In this recipe, we will focus on the latter.