internal static void DStreamTextFileSample() { count = 0; string directory = SparkCLRSamples.Configuration.SampleDataLocation; string checkpointPath = Path.Combine(directory, "checkpoint"); SparkContext sc = SparkCLRSamples.SparkContext; var b = sc.Broadcast<int>(0); StreamingContext ssc = StreamingContext.GetOrCreate(checkpointPath, () => { StreamingContext context = new StreamingContext(sc, 2); context.Checkpoint(checkpointPath); var lines = context.TextFileStream(Path.Combine(directory, "test")); lines = context.Union(lines, lines); var words = lines.FlatMap(l => l.Split(' ')); var pairs = words.Map(w => new KeyValuePair<string, int>(w, 1)); // since operations like ReduceByKey, Join and UpdateStateByKey are // separate dstream transformations defined in CSharpDStream.scala // an extra CSharpRDD is introduced in between these operations var wordCounts = pairs.ReduceByKey((x, y) => x + y); var join = wordCounts.Window(2, 2).Join(wordCounts, 2); var state = join.UpdateStateByKey<string, Tuple<int, int>, int>(new UpdateStateHelper(b).Execute); state.ForeachRDD((time, rdd) => { // there's chance rdd.Take conflicts with ssc.Stop if (stopFileServer) return; object[] taken = rdd.Take(10); Console.WriteLine("-------------------------------------------"); Console.WriteLine("Time: {0}", time); Console.WriteLine("-------------------------------------------"); foreach (object record in taken) { Console.WriteLine(record); var countByWord = (KeyValuePair<string, int>)record; Assert.AreEqual(countByWord.Value, countByWord.Key == "The" || countByWord.Key == "lazy" || countByWord.Key == "dog" ? 92 : 88); } Console.WriteLine(); stopFileServer = true; }); return context; }); StartFileServer(ssc, directory, "words.txt"); ssc.Start(); ssc.AwaitTermination(); }
public void TestStreamingContext() { var ssc = new StreamingContext(new SparkContext("", ""), 1); Assert.IsNotNull((ssc.streamingContextProxy as MockStreamingContextProxy)); ssc.Start(); ssc.Remember(1); ssc.Checkpoint(Path.GetTempPath()); var textFile = ssc.TextFileStream(Path.GetTempPath()); Assert.IsNotNull(textFile.DStreamProxy); var socketStream = ssc.SocketTextStream(IPAddress.Loopback.ToString(), 12345); Assert.IsNotNull(socketStream.DStreamProxy); var kafkaStream = KafkaUtils.CreateStream(ssc, IPAddress.Loopback + ":2181", "testGroupId", new Dictionary<string, int> { { "testTopic1", 1 } }, new Dictionary<string, string>()); Assert.IsNotNull(kafkaStream.DStreamProxy); var directKafkaStream = KafkaUtils.CreateDirectStream(ssc, new List<string> { "testTopic2" }, new Dictionary<string, string>(), new Dictionary<string, long>()); Assert.IsNotNull(directKafkaStream.DStreamProxy); var directKafkaStreamWithRepartition = KafkaUtils.CreateDirectStreamWithRepartition(ssc, new List<string> { "testTopic3" }, new Dictionary<string, string>(), new Dictionary<string, long>(), 10); Assert.IsNotNull(directKafkaStreamWithRepartition.DStreamProxy); var union = ssc.Union(textFile, socketStream); Assert.IsNotNull(union.DStreamProxy); ssc.AwaitTermination(); ssc.Stop(); }
public void TestStreamingContext() { var ssc = new StreamingContext(new SparkContext("", ""), 1000); Assert.IsNotNull((ssc.streamingContextProxy as MockStreamingContextProxy)); ssc.Start(); ssc.Remember(1000); ssc.Checkpoint(Path.GetTempPath()); var textFile = ssc.TextFileStream(Path.GetTempPath()); Assert.IsNotNull(textFile.DStreamProxy); var socketStream = ssc.SocketTextStream("127.0.0.1", 12345); Assert.IsNotNull(socketStream.DStreamProxy); var kafkaStream = ssc.KafkaStream("127.0.0.1:2181", "testGroupId", new Dictionary<string, int> { { "testTopic1", 1 } }, new Dictionary<string, string>()); Assert.IsNotNull(kafkaStream.DStreamProxy); var directKafkaStream = ssc.DirectKafkaStream(new List<string> { "testTopic2" }, new Dictionary<string, string>(), new Dictionary<string, long>()); Assert.IsNotNull(directKafkaStream.DStreamProxy); var union = ssc.Union(textFile, socketStream); Assert.IsNotNull(union.DStreamProxy); ssc.AwaitTermination(); ssc.Stop(); }
internal static void DStreamTextFileSamples() { count = 0; string directory = SparkCLRSamples.Configuration.SampleDataLocation; string checkpointPath = Path.Combine(directory, "checkpoint"); StreamingContext ssc = StreamingContext.GetOrCreate(checkpointPath, () => { SparkContext sc = SparkCLRSamples.SparkContext; StreamingContext context = new StreamingContext(sc, 2000); context.Checkpoint(checkpointPath); var lines = context.TextFileStream(Path.Combine(directory, "test")); lines = context.Union(lines, lines); var words = lines.FlatMap(l => l.Split(' ')); var pairs = words.Map(w => new KeyValuePair<string, int>(w, 1)); // since operations like ReduceByKey, Join and UpdateStateByKey are // separate dstream transformations defined in CSharpDStream.scala // an extra CSharpRDD is introduced in between these operations var wordCounts = pairs.ReduceByKey((x, y) => x + y); var join = wordCounts.Join(wordCounts, 2); var state = join.UpdateStateByKey<string, Tuple<int, int>, int>((vs, s) => vs.Sum(x => x.Item1 + x.Item2) + s); state.ForeachRDD((time, rdd) => { // there's chance rdd.Take conflicts with ssc.Stop if (stopFileServer) return; object[] taken = rdd.Take(10); Console.WriteLine("-------------------------------------------"); Console.WriteLine("Time: {0}", time); Console.WriteLine("-------------------------------------------"); foreach (object record in taken) { Console.WriteLine(record); } Console.WriteLine(); stopFileServer = count++ > 100; }); return context; }); ssc.Start(); StartFileServer(directory, "words.txt", 100); ssc.AwaitTermination(); ssc.Stop(); }
public void TestStreamingAwaitTimeout() { var ssc = new StreamingContext(new SparkContext("", ""), 1000L); Assert.IsNotNull((ssc.streamingContextProxy as MockStreamingContextProxy)); ssc.Start(); ssc.Remember(1000L); ssc.Checkpoint(Path.GetTempPath()); var textFile = ssc.TextFileStream(Path.GetTempPath()); Assert.IsNotNull(textFile.DStreamProxy); var socketStream = ssc.SocketTextStream(IPAddress.Loopback.ToString(), 12345); Assert.IsNotNull(socketStream.DStreamProxy); var union = ssc.Union(textFile, socketStream); Assert.IsNotNull(union.DStreamProxy); ssc.AwaitTerminationOrTimeout(3000); ssc.Stop(); }
internal static void DStreamMapWithStateSample() { string directory = SparkCLRSamples.Configuration.SampleDataLocation; string checkpointPath = Path.Combine(directory, "checkpoint"); StreamingContext ssc = StreamingContext.GetOrCreate(checkpointPath, () => { SparkContext sc = SparkCLRSamples.SparkContext; StreamingContext context = new StreamingContext(sc, 10000); context.Checkpoint(checkpointPath); var lines = context.TextFileStream(Path.Combine(directory, "test1")); lines = context.Union(lines, lines); var words = lines.FlatMap(l => l.Split(' ')); var pairs = words.Map(w => new KeyValuePair<string, int>(w, 1)); var wordCounts = pairs.ReduceByKey((x, y) => x + y); var initialState = sc.Parallelize(new[] { new KeyValuePair<string, int>("NOT_A_WORD", 1024), new KeyValuePair<string, int>("dog", 10000), }, 1); var stateSpec = new StateSpec<string, int, int, KeyValuePair<string, int>>((word, count, state) => { if (state.IsTimingOut()) { Console.WriteLine("Found timing out word: {0}", word); return new KeyValuePair<string, int>(word, state.Get()); } var sum = 0; if (state.Exists()) { sum = state.Get(); } state.Update(sum + count); Console.WriteLine("word: {0}, count: {1}", word, sum + count); return new KeyValuePair<string, int>(word, sum + count); }).NumPartitions(1).InitialState(initialState).Timeout(TimeSpan.FromSeconds(30)); var snapshots = wordCounts.MapWithState(stateSpec).StateSnapshots(); snapshots.ForeachRDD((double time, RDD<dynamic> rdd) => { Console.WriteLine("-------------------------------------------"); Console.WriteLine("Snapshots @ Time: {0}", time); Console.WriteLine("-------------------------------------------"); foreach (KeyValuePair<string, int> record in rdd.Collect()) { Console.WriteLine("[{0}, {1}]", record.Key, record.Value); } Console.WriteLine(); }); return context; }); ssc.Start(); StartFileServer(directory, "words.txt", 100); ssc.AwaitTermination(); ssc.Stop(); }
public void TestStreamingContext() { var ssc = new StreamingContext(new SparkContext("", ""), 1000L); Assert.IsNotNull((ssc.streamingContextProxy as MockStreamingContextProxy)); ssc.Start(); ssc.Remember(1000L); ssc.Checkpoint(Path.GetTempPath()); var textFile = ssc.TextFileStream(Path.GetTempPath()); Assert.IsNotNull(textFile.DStreamProxy); var socketStream = ssc.SocketTextStream(IPAddress.Loopback.ToString(), 12345); Assert.IsNotNull(socketStream.DStreamProxy); var kafkaStream = KafkaUtils.CreateStream(ssc, IPAddress.Loopback + ":2181", "testGroupId", new Dictionary<string, int> { { "testTopic1", 1 } }, new Dictionary<string, string>()); Assert.IsNotNull(kafkaStream.DStreamProxy); var directKafkaStream = KafkaUtils.CreateDirectStream(ssc, new List<string> { "testTopic2" }, new Dictionary<string, string>(), new Dictionary<string, long>()); Assert.IsNotNull(directKafkaStream.DStreamProxy); ssc.SparkContext.SparkConf.Set("spark.mobius.streaming.kafka.numPartitions.testTopic3", "10"); var directKafkaStreamWithRepartition = KafkaUtils.CreateDirectStream(ssc, new List<string> { "testTopic3" }, new Dictionary<string, string>(), new Dictionary<string, long>()); Assert.IsNotNull(directKafkaStreamWithRepartition.DStreamProxy); var directKafkaStreamWithRepartitionAndReadFunc = KafkaUtils.CreateDirectStream( ssc, new List<string> { "testTopic3" }, new Dictionary<string, string>(), new Dictionary<string, long>(), (int pid, IEnumerable<KeyValuePair<byte[], byte[]>> input) => { return input; }); Assert.IsNotNull(directKafkaStreamWithRepartitionAndReadFunc); ssc.SparkContext.SparkConf.Set("spark.mobius.streaming.kafka.numReceivers", "10"); var directKafkaReceiver = KafkaUtils.CreateDirectStream( ssc, new List<string> { "testTopic3" }, new Dictionary<string, string>(), new Dictionary<string, long>(), (int pid, IEnumerable<KeyValuePair<byte[], byte[]>> input) => { return input; }); Assert.IsNotNull(directKafkaReceiver.DStreamProxy); var union = ssc.Union(textFile, socketStream); Assert.IsNotNull(union.DStreamProxy); ssc.AwaitTermination(); ssc.Stop(); }