org.kitesdk.data.crunch
Class CrunchDatasets

java.lang.Object
  org.kitesdk.data.crunch.CrunchDatasets

@Beta
public class CrunchDatasets
extends Object

A helper class for exposing a filesystem-based dataset as a Crunch ReadableSource or Target.

Constructor Summary

CrunchDatasets()

Method Summary

static <E extends IndexedRecord> ReadableSource<E>	asSource(Dataset<E> dataset, Class<E> type) Expose the given Dataset as a Crunch ReadableSource.
static Target	asTarget(Dataset dataset) Expose the given Dataset as a Crunch Target.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

CrunchDatasets

public CrunchDatasets()

Method Detail

asSource

public static <E extends IndexedRecord> ReadableSource<E> asSource(Dataset<E> dataset,
                                                                   Class<E> type)

Expose the given Dataset as a Crunch ReadableSource. Only the FileSystem Dataset implementation is supported and the file format must be Formats.PARQUET or Formats.AVRO.

Type Parameters:

E - the type of entity produced by the source

Parameters:

dataset - the dataset to read from

type - the Java type of the entities in the dataset

Returns:

the ReadableSource, or null if the dataset is not filesystem-based.

asTarget

public static Target asTarget(Dataset dataset)

Expose the given Dataset as a Crunch Target. Only the FileSystem Dataset implementation is supported and the file format must be Formats.PARQUET or Formats.AVRO. In addition, the given Dataset must not be partitioned, or must be a leaf partition in the partition hierarchy. The Target returned by this method will not write to sub-partitions.

Parameters:

dataset - the dataset to write to

Returns:

the Target, or null if the dataset is not filesystem-based.