org.apache.beam.sdk.coders

Class AvroCoder<T>

java.lang.Object

org.apache.beam.sdk.coders.StandardCoder<T>

org.apache.beam.sdk.coders.AvroCoder<T>

Type Parameters:

T - the type of elements handled by this coder

All Implemented Interfaces:

Serializable, Coder<T>

public class AvroCoder<T>
extends StandardCoder<T>

A Coder using Avro binary format.

Each instance of AvroCoder<T> encapsulates an Avro schema for objects of type T.

The Avro schema may be provided explicitly via of(Class, Schema) or omitted via of(Class), in which case it will be inferred using Avro's ReflectData.

For complete details about schema generation and how it can be controlled please see the org.apache.avro.reflect package. Only concrete classes with a no-argument constructor can be mapped to Avro records. All inherited fields that are not static or transient are included. Fields are not permitted to be null unless annotated by Nullable or a Union schema containing "null".

To use, specify the Coder type on a PCollection:

 
 PCollection<MyCustomElement> records =
     input.apply(...)
          .setCoder(AvroCoder.of(MyCustomElement.class);
 
 

or annotate the element class using @DefaultCoder.

  @DefaultCoder(AvroCoder.class)
 public class MyCustomElement {
   ...
 }
 

The implementation attempts to determine if the Avro encoding of the given type will satisfy the criteria of Coder.verifyDeterministic() by inspecting both the type and the Schema provided or generated by Avro. Only coders that are deterministic can be used in GroupByKey operations.

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from interface org.apache.beam.sdk.coders.Coder

Coder.Context, Coder.NonDeterministicException

Field Summary

Fields

Modifier and Type	Field and Description
static CoderProvider	PROVIDER

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	AvroCoder(Class<T> type, Schema schema)

Method Summary

Modifier and Type	Method and Description
org.apache.beam.sdk.util.CloudObject	asCloudObject() Returns the CloudObject that represents this Coder.
DatumReader<T>	createDatumReader() Deprecated. For AvroCoder internal use only.
DatumWriter<T>	createDatumWriter() Deprecated. For AvroCoder internal use only.
T	decode(InputStream inStream, Coder.Context context) Decodes a value of type T from the given input stream in the given context.
void	encode(T value, OutputStream outStream, Coder.Context context) Encodes the given value of type T onto the given output stream in the given context.
List<? extends Coder<?>>	getCoderArguments() If this is a Coder for a parameterized type, returns the list of Coders being used for each of the parameters, or returns null if this cannot be done or this is not a parameterized type.
String	getEncodingId() The encoding identifier is designed to support evolution as per the design of Avro In order to use this class effectively, carefully read the Avro documentation at Schema Resolution to ensure that the old and new schema match.
Schema	getSchema() Returns the schema used by this coder.
Class<T>	getType() Returns the type this coder encodes/decodes.
static <T> AvroCoder<T>	of(Class<T> clazz) Returns an AvroCoder instance for the provided element class.
static <T> AvroCoder<T>	of(Class<T> type, Schema schema) Returns an AvroCoder instance for the provided element type using the provided Avro schema.
static AvroCoder<GenericRecord>	of(Schema schema) Returns an AvroCoder instance for the Avro schema.
static AvroCoder<?>	of(String classType, String schema)
static <T> AvroCoder<T>	of(TypeDescriptor<T> type) Returns an AvroCoder instance for the provided element type.
void	verifyDeterministic() Throw Coder.NonDeterministicException if the coding is not deterministic.

Methods inherited from class org.apache.beam.sdk.coders.StandardCoder

consistentWithEquals, equals, getAllowedEncodings, getComponents, getEncodedElementByteSize, hashCode, isRegisterByteSizeObserverCheap, registerByteSizeObserver, structuralValue, toString, verifyDeterministic, verifyDeterministic

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Field Detail

PROVIDER

public static final CoderProvider PROVIDER

Constructor Detail

AvroCoder

protected AvroCoder(Class<T> type,
                    Schema schema)

Method Detail

of

public static <T> AvroCoder<T> of(TypeDescriptor<T> type)

Returns an AvroCoder instance for the provided element type.

Type Parameters:

T - the element type

of

public static <T> AvroCoder<T> of(Class<T> clazz)

Returns an AvroCoder instance for the provided element class.

Type Parameters:

T - the element type

of

public static AvroCoder<GenericRecord> of(Schema schema)

Returns an AvroCoder instance for the Avro schema. The implicit type is GenericRecord.

of

public static <T> AvroCoder<T> of(Class<T> type,
                                  Schema schema)

Returns an AvroCoder instance for the provided element type using the provided Avro schema.

If the type argument is GenericRecord, the schema may be arbitrary. Otherwise, the schema must correspond to the type provided.

Type Parameters:

T - the element type

of

public static AvroCoder<?> of(String classType,
                              String schema)
                       throws ClassNotFoundException

Throws:

ClassNotFoundException

getEncodingId

public String getEncodingId()

The encoding identifier is designed to support evolution as per the design of Avro In order to use this class effectively, carefully read the Avro documentation at Schema Resolution to ensure that the old and new schema match.

In particular, this encoding identifier is guaranteed to be the same for AvroCoder instances of the same principal class, and otherwise distinct. The schema is not included in the identifier.

When modifying a class to be encoded as Avro, here are some guidelines; see the above link for greater detail.

• Avoid changing field names.
• Never remove a required field.
• Only add optional fields, with sensible defaults.
• When changing the type of a field, consult the Avro documentation to ensure the new and old types are interchangeable.

Code consuming this message class should be prepared to support all versions of the class until it is certain that no remaining serialized instances exist.

If backwards incompatible changes must be made, the best recourse is to change the name of your class.

Specified by:

getEncodingId in interface Coder<T>

Overrides:

getEncodingId in class StandardCoder<T>

getType

public Class<T> getType()

Returns the type this coder encodes/decodes.

encode

public void encode(T value,
                   OutputStream outStream,
                   Coder.Context context)
            throws IOException

Description copied from interface: Coder

Encodes the given value of type T onto the given output stream in the given context.

Throws:

IOException - if writing to the OutputStream fails for some reason

CoderException - if the value could not be encoded for some reason

decode

public T decode(InputStream inStream,
                Coder.Context context)
         throws IOException

Description copied from interface: Coder

Decodes a value of type T from the given input stream in the given context. Returns the decoded value.

Throws:

IOException - if reading from the InputStream fails for some reason

CoderException - if the value could not be decoded for some reason

getCoderArguments

public List<? extends Coder<?>> getCoderArguments()

Description copied from interface: Coder

If this is a Coder for a parameterized type, returns the list of Coders being used for each of the parameters, or returns null if this cannot be done or this is not a parameterized type.

asCloudObject

public org.apache.beam.sdk.util.CloudObject asCloudObject()

Description copied from interface: Coder

Returns the CloudObject that represents this Coder.

Specified by:

asCloudObject in interface Coder<T>

Overrides:

asCloudObject in class StandardCoder<T>

verifyDeterministic

public void verifyDeterministic()
                         throws Coder.NonDeterministicException

Description copied from interface: Coder

Throw Coder.NonDeterministicException if the coding is not deterministic.

In order for a Coder to be considered deterministic, the following must be true:

• two values that compare as equal (via Object.equals() or Comparable.compareTo(), if supported) have the same encoding.
• the Coder always produces a canonical encoding, which is the same for an instance of an object even if produced on different computers at different times.

Throws:

NonDeterministicException - when the type may not be deterministically encoded using the given Schema, the directBinaryEncoder, and the ReflectDatumWriter or GenericDatumWriter.

Coder.NonDeterministicException - if this coder is not deterministic.

createDatumReader

@Deprecated
public DatumReader<T> createDatumReader()

Deprecated. For AvroCoder internal use only.

Returns a new DatumReader that can be used to read from an Avro file directly. Assumes the schema used to read is the same as the schema that was used when writing.

createDatumWriter

@Deprecated
public DatumWriter<T> createDatumWriter()

Deprecated. For AvroCoder internal use only.

Returns a new DatumWriter that can be used to write to an Avro file directly.

getSchema

public Schema getSchema()

Returns the schema used by this coder.