将一段内存变量映射为一个P类型实例 Args: dataset (object): 任意类型的内存变量 options: serde: 设置dataset的serde对象

Apply a transform on itself, p.apply(transform, *args) is equivalent to p.transform(*args) Args: transform (function): tr

对包含的元素进行变换，等同于 ``transforms.map(self, fn, *side_inputs, **options)`` Args: fn (function): 变换函数 *side_inputs: 参

Data size of this node

Get runtime value from this PType, equivalent to ``pipeline.get(self)``. Returns: object: runtime value

对Value进行一个变换 Args: transform (callable): 作用在Value上的变换函数 *args (object): 变换所需要的参数列表 **options: 可配置选项

与其他PCollection做连接操作，等同于 :func:`bigflow.transforms.join(self, other, *others) <bigflow.transforms.join>` Args:

find all units in all the successors of this visitor very high costs

no comment

将外部存储的数据映射为一个PCollection，并在运行时读取数据 Args: source (Source): 表示外部存储的Source实例 Returns: PCollection: 读取结果

Return the pipeline this PType belongs to Returns: Pipeline: pipeline

将一个PCollection映射为外部存储数据，并在运行期写到该外部存储 Args: pcollection (PCollection): 要写出的PCollection target (Target): 表示外部存储的

return the serde who is used to create this node

Return Protobuf message of a LogicalPlan Node or Scope :param node_or_scope: Node or Scope :return: Protobuf message

tests _test_this: construct test dict Args: need_sorted: for list assert, default: we sort before assert

对包含的元素进行一对多变换，等同于 ``transforms.map(self, fn, *side_inputs, **options)`` Args: fn (function): 变换函数 *side_inputs:

inner

将一个P类型表示的数据汇聚为内存变量，相当于调用pvalue.get()。改方法隐式调用pvalue.cache() 并立即触发Pipeline.run() Args: pvalue (PType): P类型实例

返回该Pipeline的默认序列化/反序列化器 Returns: Objector: 序列化/反序列化器，用户不应当修改此objector