spark graphx: creat an graph from csv

创建图模型的方法有很多，除了GraphLoader.edgeListFile()外，还可以通过Graph.from.EdgeTuples方法来创建图。参考：

http://stackoverflow.com/questions/32396477/how-to-create-a-graph-from-a-csv-file-using-graph-fromedgetuples-in-spark-scala/32477121#32477121

import org.apache.spark.SparkConf
import org.apache.spark.SparkContext
import org.apache.spark.graphx.GraphLoader
import scala.util.MurmurHash
import org.apache.spark.graphx.Graph
import org.apache.spark.rdd.RDD
import org.apache.spark.graphx.VertexId

object GraphFromFile {
  def main(args: Array[String]) {

    //create SparkContext
    val sparkConf = new SparkConf().setAppName("GraphFromFile").setMaster("local[*]")
    val sc = new SparkContext(sparkConf)

    // read your file
    /*suppose your data is like 
    v1 v3
    v2 v1
    v3 v4
    v4 v2
    v5 v3
    */
    val file = sc.textFile("src/main/resources/textFile1.csv");

    // create edge RDD of type RDD[(VertexId, VertexId)]
    val edgesRDD: RDD[(VertexId, VertexId)] = file.map(line => line.split(" "))
      .map(line =>
        (MurmurHash.stringHash(line(0).toString), MurmurHash.stringHash(line(1).toString)))

    // create a graph 
    val graph = Graph.fromEdgeTuples(edgesRDD, 1)

    // you can see your graph 
    graph.triplets.collect.foreach(println)

  }
}

MurmurHash.stringHash is used because file contains vertex in form of String . If its of Numeric type then it wont be required 

操作：1. 查看节点数目graph.vertices.count.2.查看边数目graph.edges.count