At 2014-08-01 11:23:49 +0800, Bin <wubin_phi...@126.com> wrote: > I am wondering what is the best way to construct a graph? > > Say I have some attributes for each user, and specific weight for each user > pair. The way I am currently doing is first read user information and edge > triple into two arrays, then use sc.parallelize to create vertexRDD and > edgeRDD, respectively. Then create the graph using Graph(vertices, edges). > > I wonder whether there is a better way to do this?
That's a perfectly fine way to construct a graph. Are you encountering a problem with it? The only suggestion I would make is to load the data using sc.textFile rather than reading into an array and calling sc.parallelize. This will avoid loading it all into the driver's memory. GraphLoader does have the slight advantage that it avoids allocating a pair per vertex, but this is unlikely to be a big cost, so it's fine to use Graph(vertices, edges) if GraphLoader isn't suitable. Ankur
