org.apache.crunch.types.orc

Class Orcs

java.lang.Object

org.apache.crunch.types.orc.Orcs

public class Orcs
extends Object

Utilities to create PTypes for ORC serialization / deserialization

Constructor Summary

Constructors

Constructor and Description
Orcs()

Method Summary

Modifier and Type	Method and Description
static PType<org.apache.hadoop.hive.ql.io.orc.OrcStruct>	orcs(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo typeInfo) Create a PType to directly use OrcStruct as the deserialized format.
static <T> PType<T>	reflects(Class<T> clazz) Create a PType which uses reflection to serialize/deserialize java POJOs to/from ORC.
static PType<TupleN>	tuples(PType... ptypes) Create a tuple-based PType.

Methods inherited from class java.lang.Object

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

Constructor Detail

Orcs

public Orcs()

Method Detail

orcs

public static final PType<org.apache.hadoop.hive.ql.io.orc.OrcStruct> orcs(org.apache.hadoop.hive.serde2.typeinfo.TypeInfo typeInfo)

Create a PType to directly use OrcStruct as the deserialized format. This is the fastest way for serialization/deserializations. However, users need to use ObjectInspectors to handle the OrcStruct. Currently, void and union types are not supported.

Parameters:

typeInfo -

Returns:

reflects

public static final <T> PType<T> reflects(Class<T> clazz)

Create a PType which uses reflection to serialize/deserialize java POJOs to/from ORC. There are some restrictions of the POJO: 1) it must have a default, no-arg constructor; 2) All of its fields must be Hive primitive types or collection types that have Hive equivalents; 3) Void and Union are not supported yet.

Parameters:

clazz -

Returns:

tuples

public static final PType<TupleN> tuples(PType... ptypes)

Create a tuple-based PType. Users can use other Crunch PTypes (such as Writables.ints(), Orcs.reflects(), Writables.pairs(), ...) to construct the PType. Currently, nulls and unions are not supported.

Parameters:

ptypes -

Returns: