PCollection and other classes for
representing data in a Pipeline.See: Description
| Interface | Description |
|---|---|
| PCollectionView<T> |
A
PCollectionView<T> is an immutable view of a PCollection
as a value of type T that can be accessed
as a side input to a ParDo transform. |
| PInput |
The interface for things that might be input to a
PTransform. |
| POutput |
The interface for things that might be output from a
PTransform. |
| PValue |
The interface for values that can be input to and output from
PTransforms. |
| Class | Description |
|---|---|
| KV<K,V> |
An immutable key/value pair.
|
| KV.OrderByKey<K extends Comparable<? super K>,V> |
A
Comparator that orders KVs by the natural ordering of their keys. |
| KV.OrderByValue<K,V extends Comparable<? super V>> |
A
Comparator that orders KVs by the natural ordering of their values. |
| PBegin | |
| PCollection<T> |
A
PCollection<T> is an immutable collection of values of type
T. |
| PCollectionList<T> |
A
PCollectionList<T> is an immutable list of homogeneously
typed PCollection<T>s. |
| PCollectionTuple |
A
PCollectionTuple is an immutable tuple of
heterogeneously-typed PCollections, "keyed" by
TupleTags. |
| PDone | |
| POutputValueBase |
A
POutputValueBase is the abstract base class of
PTransform outputs. |
| PValueBase |
A
PValueBase is an abstract base class that provides
sensible default implementations for methods of PValue. |
| TimestampedValue<V> |
An immutable pair of a value and a timestamp.
|
| TimestampedValue.TimestampedValueCoder<T> |
A
Coder for TimestampedValue. |
| TupleTag<V> |
A
TupleTag is a typed tag to use as the key of a
heterogeneously typed tuple, like PCollectionTuple. |
| TupleTagList |
A
TupleTagList is an immutable list of heterogeneously
typed TupleTags. |
| TypeDescriptor<T> |
A description of a Java type, including actual generic parameters where possible.
|
| TypeDescriptors |
A utility class containing the Java primitives for
TypeDescriptor equivalents. |
| TypedPValue<T> |
A
TypedPValue<T> is the abstract base class of things that
store some number of values of type T. |
| Enum | Description |
|---|---|
| PCollection.IsBounded |
The enumeration of cases for whether a
PCollection is bounded. |
PCollection and other classes for
representing data in a Pipeline.
In particular, see these collection abstractions:
PCollection - an immutable collection of
values of type T and the main representation for data in Dataflow.PCollectionView - an immutable view of a
PCollection that can be accessed as a
side input of a ParDo
PTransform.PCollectionTuple - a heterogeneous tuple of
PCollections
used in cases where a PTransform takes
or returns multiple
PCollections.PCollectionList - a homogeneous list of
PCollections used, for example,
as input to Flatten.And these classes for individual values play particular roles in Dataflow:
KV - a key/value pair that is used by
keyed transforms, most notably GroupByKey.
TimestampedValue - a timestamp/value pair
that is used for windowing and handling out-of-order data in streaming execution.For further details, see the documentation for each class in this package.