io.druid.data.input.schemarepo

Interface SubjectAndIdConverter<SUBJECT,ID>

All Known Implementing Classes:

Avro1124SubjectAndIdConverter

public interface SubjectAndIdConverter<SUBJECT,ID>

Schema Repository is a registry service, you can register a string schema which gives back an schema id for it, or lookup the schema with the schema id.

In order to get the "latest" schema or handle compatibility enforcement on changes there has to be some way to group a set of schemas together and reason about the ordering of changes over these. Subject is introduced as the formal notion of group, defined as an ordered collection of mutually compatible schemas, according to Scott Carey on AVRO-1124.

So you can register an string schema to a specific subject, get an schema id, and then query the schema using the subject and schema id pair. Working with Kafka and Avro, it's intuitive that using Kafka topic as subject name and an incrementing integer as schema id, serialize and attach them to the message payload, or extract and deserialize from message payload, which is implemented as Avro1124SubjectAndIdConverter.

You can implement your own SubjectAndIdConverter based on your scenario, such as using canonical name of avro schema as subject name and incrementing short integer which serialized using varint.

Method Summary

Modifier and Type	Method and Description
org.schemarepo.api.converter.Converter<ID>	getIdConverter()
Pair<SUBJECT,ID>	getSubjectAndId(ByteBuffer payload)
org.schemarepo.api.converter.Converter<SUBJECT>	getSubjectConverter()
void	putSubjectAndId(SUBJECT subject, ID id, ByteBuffer payload)

Method Detail

getSubjectAndId

Pair<SUBJECT,ID> getSubjectAndId(ByteBuffer payload)

putSubjectAndId

void putSubjectAndId(SUBJECT subject,
                     ID id,
                     ByteBuffer payload)

getSubjectConverter

org.schemarepo.api.converter.Converter<SUBJECT> getSubjectConverter()

getIdConverter

org.schemarepo.api.converter.Converter<ID> getIdConverter()