flink Transitive Closure算法,实现寻找新的可达路径
flink 使用Transitive Closure算法实现可达路径查找。
1、Transitive Closure是翻译闭包传递?我觉得直译不准确,意译应该是传递特性直至特性关闭,也符合本例中传递路径,寻找路径可达,直到可达路径不存在(即关闭)。
2、代码很简单,里面有些概念直指核心原理,详细看注释。
/** * @Author: xu.dm * @Date: 2019/7/3 11:41 * @Version: 1.0 * @Description: 传递闭包算法,本例中就是根据成对路径,查找和生成新的可达路径 * 例如:1-2,2-4这两对数据,可以得出新的可达路径1-4。 * * 迭代算法步骤: * 1、获取成对数据集edges,里面包括路径对,比如 1->2,2->4,2->5等,如果是无向边,还可以反转数据集union之前的数据。本例按有向边处理 * 2、生成迭代头paths可迭代数据集 * 3、用paths和原始数据集edges做join连接,找出头尾相连的数据nextPaths,即类似1->2,2->4这种,然后生成新的路径1->4。 * 4、新的路径集nextPaths和迭代头数据集paths进行并集操作,即union操作,生成新的nextPaths,这个时候它包含了新旧两种数据 * 在这里总是nextPaths>=paths * 5、去重操作,第一次迭代不会重复,但是第二次迭代开始就会有重复数据,通过groupBy全字段,去分组第一条即可达到去重效果 * 6、以上核心迭代体完成,后面需要形成迭代闭环,确定迭代退出条件 * 7、退出原理:每次迭代完成后,需要检查是否新的路径产生,如果没有则表示迭代可以结束 * 8、可达寻路步骤完成后,通过对比nextPaths和paths,如果nextPaths>paths,表示有新路径生成,需要继续迭代,直到nextPaths=paths * 9、这里有一个迭代重要的概念,paths和nextPaths是通过迭代闭环不断更新的 * 10、本例中迭代头和迭代尾的数据流向:paths->nextPaths->paths. * 11、本例通过bulk迭代方式实现了delta迭代的效果 **/ public class TransitiveClosureNaive { public static void main(String args[]) throws Exception { // Checking input parameters final ParameterTool params = ParameterTool.fromArgs(args); // set up execution environment ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment(); // make parameters available in the web interface env.getConfig().setGlobalJobParameters(params); final int maxIterations = params.getInt("iterations", 10); DataSet<Tuple2<Long, Long>> edges; if(params.has("edges")){ edges = env.readCsvFile(params.get("edges")).fieldDelimiter(" ").types(Long.class, Long.class); }else { System.out.println("Executing TransitiveClosureNaive example with default edges data set."); System.out.println("Use --edges to specify file input."); edges = ConnectedComponentsData.getDefaultEdgeDataSet(env); } IterativeDataSet<Tuple2<Long,Long>> paths = edges.iterate(maxIterations); DataSet<Tuple2<Long,Long>> nextPaths = paths .join(edges) .where(1) .equalTo(0) .with(new JoinFunction<Tuple2<Long, Long>, Tuple2<Long, Long>, Tuple2<Long, Long>>() { /** left: Path (z,x) - 通过z可达x right: Edge (x,y) - 通过x可达y out: Path (z,y) - 最终输出z可达y */ @Override public Tuple2<Long, Long> join(Tuple2<Long, Long> left, Tuple2<Long, Long> right) throws Exception { return new Tuple2<>(left.f0,right.f1); } }) //类似withForwardedFieldsFirst这种无损转发语义声明,是可选项,有助于提高flink优化器生成更高效的执行计划 //转发第一个输入Tuple2<Long, Long>中的第一个字段,转发第二个输入Tuple2<Long, Long>中的第二个字段 .withForwardedFieldsFirst("0").withForwardedFieldsSecond("1") //合并原有的路径 .union(paths) //这里的groupBy两个fields是打算给reduceGroup去重使用 .groupBy(0,1) .reduceGroup(new GroupReduceFunction<Tuple2<Long, Long>, Tuple2<Long, Long>>() { @Override public void reduce(Iterable<Tuple2<Long, Long>> values, Collector<Tuple2<Long, Long>> out) throws Exception { out.collect(values.iterator().next()); } }) .withForwardedFields("0;1"); //对比paths以及新生成的nextPaths,获取nextPaths中比paths多的路径 //从上面的算子可以得知,nextPaths总是大于或等于paths DataSet<Tuple2<Long,Long>> newPaths = paths .coGroup(nextPaths) .where(0).equalTo(0) .with(new CoGroupFunction<Tuple2<Long, Long>, Tuple2<Long, Long>, Tuple2<Long, Long>>() { Set<Tuple2<Long, Long>> prevSet = new HashSet<>(); @Override public void coGroup(Iterable<Tuple2<Long, Long>> prevPaths, Iterable<Tuple2<Long, Long>> nextPaths, Collector<Tuple2<Long, Long>> out) throws Exception { for(Tuple2<Long,Long> prev:prevPaths){ prevSet.add(prev); } //检查有没有新的数据产生,如果有就继续迭代,否则迭代终止 for(Tuple2<Long,Long> next:nextPaths){ if(!prevSet.contains(next)){ out.collect(next); } } } }).withForwardedFieldsFirst("0").withForwardedFieldsSecond("0"); //迭代尾,在这里形成闭环,nextPaths是反馈通道,nextPaths数据集被重新传递到迭代头paths里,然后通过迭代算子不断执行。 //newPaths为空或者迭代达到最大次数,迭代结束。newPaths这里表示是否有新的路径。 //数据集迭代环:paths->nextPaths->paths DataSet<Tuple2<Long, Long>> transitiveClosure = paths.closeWith(nextPaths, newPaths); // emit result if (params.has("output")) { transitiveClosure.writeAsCsv(params.get("output"), "\n", " "); // execute program explicitly, because file sinks are lazy env.execute("Transitive Closure Example"); } else { System.out.println("Printing result to stdout. Use --output to specify output path."); transitiveClosure.print(); } } }
3、原始数据
public class ConnectedComponentsData { public static final long[] VERTICES = new long[] { 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16}; public static DataSet<Long> getDefaultVertexDataSet(ExecutionEnvironment env) { List<Long> verticesList = new LinkedList<Long>(); for (long vertexId : VERTICES) { verticesList.add(vertexId); } return env.fromCollection(verticesList); } public static final Object[][] EDGES = new Object[][] { new Object[]{1L, 2L}, new Object[]{2L, 3L}, new Object[]{2L, 4L}, new Object[]{3L, 5L}, new Object[]{6L, 7L}, new Object[]{8L, 9L}, new Object[]{8L, 10L}, new Object[]{5L, 11L}, new Object[]{11L, 12L}, new Object[]{10L, 13L}, new Object[]{9L, 14L}, new Object[]{13L, 14L}, new Object[]{1L, 15L}, new Object[]{16L, 1L} }; public static DataSet<Tuple2<Long, Long>> getDefaultEdgeDataSet(ExecutionEnvironment env) { List<Tuple2<Long, Long>> edgeList = new LinkedList<Tuple2<Long, Long>>(); for (Object[] edge : EDGES) { edgeList.add(new Tuple2<Long, Long>((Long) edge[0], (Long) edge[1])); } return env.fromCollection(edgeList); } }