本文整理了Java中cascading.tuple.Tuple类的一些代码示例，展示了Tuple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。Tuple类的具体详情如下：
包路径：cascading.tuple.Tuple
类名称：Tuple

Tuple介绍

[英]A Tuple represents a set of values. Consider a Tuple the same as a database record where every value is a column in that table.

A "tuple stream" is a set of Tuple instances passed consecutively through a Pipe assembly.

Tuples work in tandem with Fields and the TupleEntry classes. A TupleEntry holds an instance of Fields and a Tuple. It allows a tuple to be accessed by its field names, and will help maintain consistent types if any are given on the Fields instance. That is, if a field is specified at an Integer, calling #set(int,Object)with a String will force the String to be coerced into a Integer instance.

For managing custom types, see the CoercibleType interface which extends Type.

Tuple instances created by user code, by default, are mutable (or modifiable). Tuple instances created by the system are immutable (or unmodifiable, tested by calling #isUnmodifiable()).

For example tuples returned by cascading.operation.FunctionCall#getArguments(), will always be unmodifiable. Thus they must be copied if they will be changed by user code or cached in the local context. See the Tuple copy constructor, or *Copy() methods on TupleEntry.

Because a Tuple can hold any Object type, it is suitable for storing custom types. But all custom types must have a serialization support per the underlying framework.

For Hadoop, a org.apache.hadoop.io.serializer.Serialization implementation must be registered with Hadoop. For further performance improvements, see the cascading.tuple.hadoop.SerializationToken Java annotation.
[中]元组表示一组值。考虑一个元组与数据库记录相同，其中每个值是该表中的一列。
“元组流”是通过管道组件连续传递的一组元组实例。
元组与字段和TupleEntry类协同工作。TupleEntry包含字段和元组的实例。它允许通过字段名访问元组，并将帮助维护字段实例上给定的一致类型。也就是说，如果在整数处指定了字段，则使用字符串调用#set（int，Object）将强制将该字符串强制转换为整数实例。
有关管理自定义类型的信息，请参见扩展类型的强制类型接口。
默认情况下，由用户代码创建的元组实例是可变的（或可修改的）。系统创建的元组实例是不可变的（或不可修改的，通过调用#isUnmodifiable（）进行测试）。
例如，通过级联返回的元组。活动FunctionCall#getArguments（）始终不可修改。因此，如果它们将被用户代码更改或缓存在本地上下文中，则必须复制它们。请参阅Tuple copy构造函数，或TupleEntry上的*copy（）方法。
因为元组可以保存任何对象类型，所以它适合存储自定义类型。但所有自定义类型都必须根据底层框架具有序列化支持。
对于Hadoop，一个组织。阿帕奇。hadoop。伊奥。序列化程序。序列化实现必须在Hadoop中注册。有关进一步的性能改进，请参阅级联。元组。hadoop。SerializationToken Java注释。

代码示例

代码示例来源：origin: LiveRamp/cascading_ext

@Override
public Tuple toTuple(Number[] values) {
 Tuple tuple = new Tuple();
 for (Number value : values) {
  tuple.add(value);
 }
 return tuple;
}

代码示例来源：origin: elastic/elasticsearch-hadoop

@Override
  public void convert(Object from, BytesArray to) {
    // expect a tuple holding one field - chararray or bytearray
    Assert.isTrue(from instanceof SinkCall,
        String.format("Unexpected object type, expecting [%s], given [%s]", SinkCall.class, from.getClass()));

    // handle common cases
    SinkCall sinkCall = (SinkCall) from;
    Tuple rawTuple = sinkCall.getOutgoingEntry().getTuple();

    if (rawTuple == null || rawTuple.isEmpty()) {
      to.bytes("{}");
      return;
    }
    Assert.isTrue(rawTuple.size() == 1, "When using JSON input, only one field is expected");

    // postpone the coercion
    Tuple tuple = CascadingUtils.coerceToString(sinkCall);
    super.convert(tuple.getObject(0), to);
  }
}

cascading.tuple.Tuple类的使用及代码示例

Tuple介绍

代码示例

相关文章

热门标签

最新文章

Tuple类方法