Hi flinksters, I'm really close to the end (at least I hope so), but I still have some issues. Writing my final loop I got this error: org.apache.flink.compiler.CompilerException: An error occurred while translating the optimized plan to a nephele JobGraph: An error occurred while translating the optimized plan to a nephele JobGraph: Error: It is currently not supported to union between dynamic and static path in an iteration. at org.apache.flink.compiler.plantranslate.NepheleJobGraphGenerator.postVisit(NepheleJobGraphGenerator.java:567) at org.apache.flink.compiler.plantranslate.NepheleJobGraphGenerator.postVisit(NepheleJobGraphGenerator.java:96) at org.apache.flink.compiler.plan.SingleInputPlanNode.accept(SingleInputPlanNode.java:202) at org.apache.flink.compiler.plan.SingleInputPlanNode.accept(SingleInputPlanNode.java:196) at org.apache.flink.compiler.plan.SingleInputPlanNode.accept(SingleInputPlanNode.java:196) I guess that I should use just the loop step set to create the next step set? Here is the code: def createPlanFullIterations(env: ExecutionEnvironment) = { val tmp = env readTextFile config.inFile map {Vector.parseFromSVMLightString (config.dimensions, _)} val X = tmp map {_._1} var residual = tmp map {_._2} val randoms = env readTextFile config.randomFile map {Vector.parseFromString(_)} val widthCandidates: DataSet[Vector] = if(config.widthCandidatesFile != null) env readTextFile config.widthCandidatesFile map {Vector.parseFromString(config.dimensions, _)} else null val emptyDataSet = env.fromCollection[Vector](Seq()) val model = emptyDataSet.iterate(config.iterations){ stepSet => val center = calcCenter(env, X, residual, randoms, -1) val x = calcWidthHeight(env, X, residual, widthCandidates, center) val width = x._1 val height = x._2 residual = residual - (getKernelVector(X, center, width).map(new RichMapFunction[Vector, Vector]{ var height: Vector = null override def open(config: Configuration) = { height = getRuntimeContext.getBroadcastVariable("height").toList.head } def map(x: Vector) = {x * height} }).withBroadcastSet(height, "height")) val centerOut = center map {x => new Vector(0, x.values)} val widthOut = width map {x => new Vector(1, x.values)} val heightOut = height map {x => new Vector(2, x.values)} val stepModel = centerOut union widthOut union height stepSet union (stepModel.map(new RichMapFunction[Vector, Vector]{ def map(x: Vector) = new Vector((getIterationRuntimeContext.getSuperstepNumber-1)*3+x.id, x.values) })) } model map { _ toString } writeAsText config.outFile } thanks! Cheers, Max |
Hi! Yes, that is a known limitation. It is sort of straightforward to resolve, but will take a bit of time. I will try and get to it until the end of the week. Stephan On Tue, Dec 9, 2014 at 11:21 AM, Maximilian Alber <[hidden email]> wrote:
|
Thanks! On Wed, Dec 10, 2014 at 4:35 PM, Stephan Ewen <[hidden email]> wrote:
|
Free forum by Nabble | Edit this page |