spark one to one shortest path using pregel graphx

Question

i tried to find shortest path from single source to n vertices using code from link

val graph: Graph[Long, Double] =
  GraphGenerators.logNormalGraph(sc, numVertices = 100).mapEdges(e => e.attr.toDouble)
val sourceId: VertexId = 42

val initialGraph = graph.mapVertices((id, _) =>
    if (id == sourceId) 0.0 else Double.PositiveInfinity)
val sssp = initialGraph.pregel(Double.PositiveInfinity)(
  (id, dist, newDist) => math.min(dist, newDist),
  triplet => {
    if (triplet.srcAttr + triplet.attr < triplet.dstAttr) {
      Iterator((triplet.dstId, triplet.srcAttr + triplet.attr))
    } else {
      Iterator.empty
    }
  },
  (a, b) => math.min(a, b)
)
println(sssp.vertices.collect.mkString("\n"))

it gives me output shortest path from 42 to N verticies. however, how to find the shortest path between single source to single destination? i.e. source=42, dest = 135 then i want to find the shortest path between them.

thanks

The way I see it there is no guarantee in which order the massages to other vertices are processed. Is that right? — Viktor Seifert, Aug 11 '17 at 07:55

score 0 · Answer 1 · answered Dec 10 '18 at 21:17

You can't find the shortest distance between two vertices otherwise i.e., without exploring the whole graph. Because, even if you stop your exploration once your message is received by the destination, nothing guarantees that it's the shortest distance. So, this algorithm is fine in terms of complexity. You should just pick the sssp value from the different destinations obtained.

spark one to one shortest path using pregel graphx

1 Answers1