Mastering Machine Learning with Spark 2.x

Building and loading graphs

In the last section, we made a lot of leeway in graph analytics and discussed an interesting retweet graph. Before we dive into more complicated operations, let's take a step back and consider other options to construct graphs with GraphX. Having completed this interlude, we will have a quick look into visualization tools and then turn to the more involved applications.

In fact, we have already seen two ways to create GraphX graphs, one was to construct the vertex and edge RDDs explicitly, ourselves, to construct a graph from it; the other one was to use Graph.fromEdges. Another very handy possibility is to load a so-called edge list file. An example of this format is the following:

1 35 34 23 21 5

So, an edge ...