parquet.hadoop.metadata

Class GlobalMetaData

java.lang.Object

parquet.hadoop.metadata.GlobalMetaData

All Implemented Interfaces:

Serializable

public class GlobalMetaData
extends Object
implements Serializable

Merged metadata when reading from multiple files. THis is to allow schema evolution

Author:

Julien Le Dem

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
GlobalMetaData(MessageType schema, Map<String,Set<String>> keyValueMetaData, Set<String> createdBy)

Method Summary

Methods

Modifier and Type	Method and Description
Set<String>	getCreatedBy()
Map<String,Set<String>>	getKeyValueMetaData()
MessageType	getSchema()
FileMetaData	merge() Will merge the metadata as if it was coming from a single file.
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

GlobalMetaData

public GlobalMetaData(MessageType schema,
              Map<String,Set<String>> keyValueMetaData,
              Set<String> createdBy)

Parameters:

schema - the union of the schemas for all the files

keyValueMetaData - the merged app specific metadata

createdBy - the description of the library that created the file

Method Detail

getSchema

public MessageType getSchema()

Returns:

the schema for the file

toString

public String toString()

Overrides:

toString in class Object

getKeyValueMetaData

public Map<String,Set<String>> getKeyValueMetaData()

Returns:

meta data for extensions

getCreatedBy

public Set<String> getCreatedBy()

Returns:

the description of the library that created the file

merge

public FileMetaData merge()

Will merge the metadata as if it was coming from a single file. (for all part files written together this will always work) If there are conflicting values an exception will be thrown

Returns:

the merged version of this