public class WithTimestamps<T> extends PTransform<PCollection<T>,PCollection<T>>
PTransform for assigning timestamps to all the elements of a PCollection.
Timestamps are used to assign Windows to elements within the
Window.into(com.google.cloud.dataflow.sdk.transforms.windowing.WindowFn)
PTransform. Assigning timestamps is useful when the input data set comes from a
Source without implicit timestamps (such as
TextIO).
name| Modifier and Type | Method and Description |
|---|---|
PCollection<T> |
apply(PCollection<T> input)
Applies this
PTransform on the given InputT, and returns its
Output. |
Duration |
getAllowedTimestampSkew()
Returns the allowed timestamp skew duration, which is the maximum
duration that timestamps can be shifted backwards from the timestamp of the input element.
|
static <T> WithTimestamps<T> |
of(SerializableFunction<T,Instant> fn)
For a
SerializableFunction fn from T to Instant, outputs a
PTransform that takes an input PCollection<T> and outputs a
PCollection<T> containing every element v in the input where
each element is output with a timestamp obtained as the result of fn.apply(v). |
WithTimestamps<T> |
withAllowedTimestampSkew(Duration allowedTimestampSkew)
Return a new WithTimestamps like this one with updated allowed timestamp skew, which is the
maximum duration that timestamps can be shifted backward.
|
getDefaultOutputCoder, getDefaultOutputCoder, getDefaultOutputCoder, getKindString, getName, toString, validatepublic static <T> WithTimestamps<T> of(SerializableFunction<T,Instant> fn)
SerializableFunction fn from T to Instant, outputs a
PTransform that takes an input PCollection<T> and outputs a
PCollection<T> containing every element v in the input where
each element is output with a timestamp obtained as the result of fn.apply(v).
If the input PCollection elements have timestamps, the output timestamp for each
element must not be before the input element's timestamp minus the value of
getAllowedTimestampSkew(). If an output timestamp is before this time, the transform
will throw an IllegalArgumentException when executed. Use
withAllowedTimestampSkew(Duration) to update the allowed skew.
Each output element will be in the same windows as the input element. If a new window based
on the new output timestamp is desired, apply a new instance of Window.into(WindowFn).
This transform will fail at execution time with a NullPointerException if for any
input element the result of fn.apply(v) is null.
Example of use in Java 8:
PCollection<Record> timestampedRecords = records.apply(
WithTimestamps.of((Record rec) -> rec.getInstant());
public WithTimestamps<T> withAllowedTimestampSkew(Duration allowedTimestampSkew)
The default value is Duration.ZERO, allowing timestamps to only be shifted into the
future. For infinite skew, use new Duration(Long.MAX_VALUE).
public Duration getAllowedTimestampSkew()
DoFn.getAllowedTimestampSkew()public PCollection<T> apply(PCollection<T> input)
PTransformPTransform on the given InputT, and returns its
Output.
Composite transforms, which are defined in terms of other transforms, should return the output of one of the composed transforms. Non-composite transforms, which do not apply any transforms internally, should return a new unbound output and register evaluators (via backend-specific registration methods).
The default implementation throws an exception. A derived class must
either implement apply, or else each runner must supply a custom
implementation via
PipelineRunner.apply(com.google.cloud.dataflow.sdk.transforms.PTransform<InputT, OutputT>, InputT).
apply in class PTransform<PCollection<T>,PCollection<T>>