本文整理了Java中cz.seznam.euphoria.core.client.dataset.Dataset类的一些代码示例，展示了Dataset类的具体用法。这些代码示例主要来源于Github/Stackoverflow/Maven等平台，是从一些精选项目中提取出来的代码，具有较强的参考意义，能在一定程度帮忙到你。Dataset类的具体详情如下：
包路径：cz.seznam.euphoria.core.client.dataset.Dataset
类名称：Dataset

Dataset介绍

[英]A dataset abstraction.
[中]数据集抽象。

代码示例

代码示例来源：origin: seznam/euphoria

public <S extends DataSink<T>> Dataset<T> persist(S dst) {
 this.wrap.persist(dst);
 return this;
}

代码示例来源：origin: seznam/euphoria

@SuppressWarnings("unchecked")
private static void applySinkTransforms(Flow flow) {
 List<Dataset<?>> outputs = flow.operators().stream()
   .filter(o -> o.output().getOutputSink() != null)
   .map(Operator::output)
   .collect(Collectors.toList());
 outputs.forEach(d -> {
  if (d.getOutputSink().prepareDataset((Dataset) d)) {
   // remove the old output sink
   d.persist(null);
  }
 });
}

代码示例来源：origin: seznam/euphoria

/**
 * Create output dataset for given operator.
 *
 * @param <IN> the type of elements of the input dataset
 * @param <OUT> the type of elements in the output dataset
 *
 * @param flow the flow to associate the output dataset with
 * @param input the input dataset the output dataset is indirectly derived from
 * @param op the operator producing the output dataset
 *
 * @return a dataset representing the output of the given operator
 */
public static <IN, OUT> Dataset<OUT> createOutputFor(
  Flow flow, Dataset<IN> input, Operator<IN, OUT> op) {
 return new OutputDataset<>(flow, op, input.isBounded());
}

代码示例来源：origin: seznam/euphoria

/**
 * Validate that no two output datasets use the same sink.
 * This is not supported, because we cannot clone the sink
 * (it can be used in client code after the flow has completed).
 */
@SuppressWarnings("unchecked")
private static void checkSinks(DAG<Operator<?, ?>> dag) {
 List<Pair<Dataset, DataSink>> outputs = dag.nodes()
   .filter(n -> n.output().getOutputSink() != null)
   .map(o -> Pair.of((Dataset) o.output(), (DataSink) o.output().getOutputSink()))
   .collect(Collectors.toList());
 Map<DataSink, Dataset> sinkDatasets = new HashMap<>();
 outputs.forEach(p -> {
  Dataset current = sinkDatasets.get(p.getSecond());
  if (current != null) {
   throw new IllegalArgumentException(
     "Operator " + current.getProducer().getName() + " and "
       + " operator " + p.getFirst().getProducer().getName()
       + " use the same sink " + p.getSecond());
  }
  sinkDatasets.put(p.getSecond(), p.getFirst());
 });
}

代码示例来源：origin: seznam/euphoria

ds.persist(HFileSink.newBuilder()
  .withConfiguration(conf)
  .withZookeeperQuorum(outputQuorum)
  .withTable(table)
  .withOutputPath(new Path(tmpDir))
  .applyIf(ds.isBounded(),
    b -> b.windowBy(GlobalWindowing.get(), w -> ""),
    b -> b.windowBy(

代码示例来源：origin: seznam/euphoria

InputOperator(Dataset<T> ds) {
 super("InputOperator", ds.getFlow());
 this.ds = ds;
 this.hints = Collections.emptySet();
}

代码示例来源：origin: seznam/euphoria

InputProvider get(Operator<?, ?> source, Operator<?, ?> target) {
 Pair<Operator<?, ?>, Operator<?, ?>> edge = Pair.of(source, target);
 InputProvider sup = materializedOutputs.get(edge);
 if (sup == null) {
  throw new IllegalArgumentException(String.format(
    "Do not have suppliers for edge %s -> %s (original producer %s )",
    source, target, source.output().getProducer()));
 }
 return sup;
}
void markRunning(Operator<?, ?> operator) {

代码示例来源：origin: seznam/euphoria

.stream()
.map(Node::get)
.filter(op -> op.output().getOutputSink() != null)
.forEach(op -> {
 final DataSink<?> sink = op.output().getOutputSink();
 sinks.add(sink);
 DataSet<?> flinkOutput =

代码示例来源：origin: seznam/euphoria

DataSource<?> raw = op.output().getSource();
if (raw.isBounded()) {
 BoundedDataSource<?> source = raw.asBounded();

代码示例来源：origin: seznam/euphoria

public <LEFT, RIGHT> ByBuilder<LEFT, RIGHT> of(Dataset<LEFT> left, Dataset<RIGHT> right) {
  if (right.getFlow() != left.getFlow()) {
   throw new IllegalArgumentException("Pass inputs from the same flow");
  }
  return new ByBuilder<>(name, left, right);
 }
}

代码示例来源：origin: seznam/euphoria

InputProvider get(Operator<?, ?> source, Operator<?, ?> target) {
 Pair<Operator<?, ?>, Operator<?, ?>> edge = Pair.of(source, target);
 InputProvider sup = materializedOutputs.get(edge);
 if (sup == null) {
  throw new IllegalArgumentException(String.format(
    "Do not have suppliers for edge %s -> %s (original producer %s )",
    source, target, source.output().getProducer()));
 }
 return sup;
}
void markRunning(Operator<?, ?> operator) {

代码示例来源：origin: seznam/euphoria

.stream()
.map(Node::get)
.filter(op -> op.output().getOutputSink() != null)
.forEach(op -> {
 final DataSink<?> sink = op.output().getOutputSink();
 sinks.add(sink);
 JavaRDD<SparkElement> sparkOutput =

代码示例来源：origin: seznam/euphoria

/**
 * Validate that no two output datasets use the same sink.
 * This is not supported, because we cannot clone the sink
 * (it can be used in client code after the flow has completed).
 */
@SuppressWarnings("unchecked")
private static void checkSinks(DAG<Operator<?, ?>> dag) {
 List<Pair<Dataset, DataSink>> outputs = dag.nodes()
   .filter(n -> n.output().getOutputSink() != null)
   .map(o -> Pair.of((Dataset) o.output(), (DataSink) o.output().getOutputSink()))
   .collect(Collectors.toList());
 Map<DataSink, Dataset> sinkDatasets = new HashMap<>();
 outputs.forEach(p -> {
  Dataset current = sinkDatasets.get(p.getSecond());
  if (current != null) {
   throw new IllegalArgumentException(
     "Operator " + current.getProducer().getName() + " and "
       + " operator " + p.getFirst().getProducer().getName()
       + " use the same sink " + p.getSecond());
  }
  sinkDatasets.put(p.getSecond(), p.getFirst());
 });
}

代码示例来源：origin: seznam/euphoria

output = createStream(op.output().getSource());
} else if (op instanceof FlatMap) {
 output = execMap((Node) node, context);

代码示例来源：origin: seznam/euphoria

/**
 * Persist given dataset into this sink via given mapper.
 * @param <T> input datatype
 * @param input the input dataset
 * @param mapper map function for transformation of input value into {@link Cell}.
 */
public <T> void persist(Dataset<T> input, UnaryFunction<T, Cell> mapper) {
 MapElements.of(input)
   .using(mapper)
   .output()
   .persist(this);
}

代码示例来源：origin: seznam/euphoria

@SuppressWarnings("unchecked")
private static void applySinkTransforms(Flow flow) {
 List<Dataset<?>> outputs = flow.operators().stream()
   .filter(o -> o.output().getOutputSink() != null)
   .map(Operator::output)
   .collect(Collectors.toList());
 outputs.forEach(d -> {
  if (d.getOutputSink().prepareDataset((Dataset) d)) {
   // remove the old output sink
   d.persist(null);
  }
 });
}

代码示例来源：origin: seznam/euphoria

public <LEFT, RIGHT> ByBuilder<LEFT, RIGHT> of(Dataset<LEFT> left, Dataset<RIGHT> right) {
  if (right.getFlow() != left.getFlow()) {
   throw new IllegalArgumentException("Pass inputs from the same flow");
  }
  return new ByBuilder<>(name, left, right);
 }
}

代码示例来源：origin: seznam/euphoria

@SuppressWarnings("unchecked")
public static boolean wantTranslateBroadcastHashJoin(Join o) {
 final ArrayList<Dataset> inputs = new ArrayList(o.listInputs());
 if (inputs.size() != 2) {
  return false;
 }
 final Dataset leftDataset = inputs.get(0);
 final Dataset rightDataset = inputs.get(1);
 return
   (o.getType() == Join.Type.LEFT && hasFitsInMemoryHint(rightDataset.getProducer()) ||
     o.getType() == Join.Type.RIGHT && hasFitsInMemoryHint(leftDataset.getProducer())
   ) && !(o.getWindowing() instanceof MergingWindowing);
}

代码示例来源：origin: seznam/euphoria

.stream()
.map(Node::get)
.filter(op -> op.output().getOutputSink() != null)
.forEach(op -> {
 final DataSink<?> sink = op.output().getOutputSink();
 sinks.add(sink);
 DataStream<?> flinkOutput =

代码示例来源：origin: seznam/euphoria

/**
 * Checks if the given {@link Flow} reads bounded inputs
 *
 * @param flow the flow to inspect
 *
 * @return {@code true} if all sources are bounded
 */
private boolean isBoundedInput(Flow flow) {
 // check if sources are bounded or not
 for (Dataset<?> ds : flow.sources()) {
  if (!ds.isBounded()) {
   return false;
  }
 }
 return true;
}