I've developed a custom source connector for an external REST service.
I get JSONs, convert them to org.apache.kafka.connect.data.Struct with manually defined schema (SchemaBuilder) and wrap all this to SourceRecord.
All of this is for one entity only, but there a dozen of them.
My new goal is to make this connector universal and parametrize the schema. The idea is to get the schema as String (json) from configs or external files and pass it to SourceRecord, but it only accepts Schema objects.
Is there any simple/good ways to convert String/json to Schema or even pass String schema directly?
There is a JSON to Avro converter, however, if you are already building a Struct/Schema combination, then you shouldn't need to do anything, as the Converter classes in Kafka Connect can handle the conversion for you
https://www.confluent.io/blog/kafka-connect-deep-dive-converters-serialization-explained/
I am working on datastore datasource for apache-spark based on spark datasource V2 api. I was able to implement using hard-coded single entity but couldn't generalize it. Either I need to infer entity schema and translate entity record into Spark Row or read entity record as json and let the user translate into scala product (datastore java client is REST based so the payload is being pulled as json). I could see "entity.properties" as json key-values from within IntelliJ debugger which includes everything I need (column name, value, type etc.) but I can't use entity.properties due to access restrictions. Appreciate any ideas.
fixed by switching to low level API https://github.com/GoogleCloudPlatform/google-cloud-datastore
full source for spark-datastore-connector https://github.com/sgireddy/spark-datastore-connector
I am trying to build a kakfa pipeline which will read JSON input data into Kafka topic.
I am using AVRO serialization with schema registry, as my schema gets changed on regular basis.
As of now GenericRecord is used to parse the schema.
But I recently came to know that avro-tools are available to read schema and generate Java classes which can be used to create Producer Code.
I am confused choose between these two options.
Can you please suggest me which one is better, as my schema gets frequently changed?
avro-tools are available to read schema and generate java classes which can be used to create Producer Code
They create specific Avro classes, not Producer code, but regarding the question. Both will work.
The way I see it
GenericRecord - Think of it as a HashMap<String, Object>. As a consumer need to know the fields to get. If, as a producer or schema creator, you are not able to send your classes as a library to your consumers, this is the essentially the best you can get. I believe you'll always be able to get the latest data, though (all possible fields can be accessed by a get("fieldname") call. See example here
SpecificRecord (what avro-tools generates) - It is just a generated class with getter methods and builder objects / setter methods. Any consumer will be able to import your producer classes as dependencies, deserialize the message, then immediately know what fields are available. You are not guaranteed to get the latest schema here - you will be "downgraded" and limited to whatever schema was used to generate those classes.
I use avro-maven-plugin to generally create the classes. Just as this example
You could also use AvroReflect to build an Avro schema from a Java class rather than the other way around. Annotations can be used on fields to set #Union or #AvroDefault settings.
Further Reading about using the Confluent Schema Registry
We are considering a serialization approach for our scala-based Akka Persistence app. We consider it likely that our persisted events will "evolve" over time, so we want to support schema evolution, and are considering Avro first.
We'd like to avoid including the full schema with every message. However, for the foreseeable future, this Akka Persistence app is the only app that will be serializing and deserializing these messages, so we don't see a need for a separate schema registry.
Checking the docs for avro and the various scala libs, I see ways to include the schema with messages, and also how to use it "schema-less" by using a schema registry, but what about the in-between case? What's the correct approach for going schema-less, but somehow including an identifier to be able to look up the correct schema (available in the local deployed codebase) for the deserialized object? Would I literally just create a schema that represents my case class, but with an additional "identifier" field for schema version, and then have some sort of in-memory map of identifier->schema at runtime?
Also, is the correct approach to have one serializer/deserialize class for each version of the schema, so it knows how to translate every version to/from the most recent version?
Finally, are there recommendations on how to unit-test schema evolutions? For instance, store a message in akka-persistence, then actually change the definition of the case class, and then kill the actor and make sure it properly evolves. (I don't see how to change the definition of the case class at runtime.)
After spending more time on this, here are the answers I came up with.
Using avro4s, you can use the default data output stream to include the schema with every serialized message. Or, you can use the binary output stream, which simply omits the schema when serializing each message. ('binary' is a bit of a misnomer here since all it does is omit the schema. In either case it is still an Array[Byte].)
Akka itself supplies a Serializer trait or a SerializerWithStringManifest trait, which will automatically include a field for a "schema identifier" in the object of whatever you serialize.
So when you create your custom serializer, you can extend the appropriate trait, define your schema identifier, and use the binary output stream. When those techniques are combined, you'll successfully be using schema-less serialization while including a schema identifier.
One common technique is to "fingerprint" your schema - treat it as a string and then calculate its digest (MD5, SHA-256, whatever). If you construct an in-memory map of fingerprint to schema, that can serve as your application's in-memory schema registry.
So then when deserializing, your incoming object will have the schema identifier of the schema that was used to serialize it (the "writer"). While deserializing, you should know the identifier of the schema to use to deserialize it (the "reader"). Avro4s supports a way for you to specify both using a builder pattern, so avro can translate the object from the old format to the new. That's how you support "schema evolution". Because of how that works, you don't need a separate serializer for each schema version. Your custom serializer will know how to evolve your objects, because that's the part that Avro gives you for free.
As for unit testing, your best bet is exploratory testing. Actually define multiple versions of a case class in your test, and multiple accompanying versions of its schema, and then explore how Avro works by writing tests that will evolve an object between different versions of that schema.
Unfortunately that won't be directly relevant to the code you are writing, because it's hard to simulate actually changing the code you are testing as you test it.
I developed a prototype that demonstrates several of these answers, and it's available on github. It uses avro, avro4s, and akka persistence. For this one, I demonstrated a changing codebase by actually changing it across commits - you'd check out commit #1, run the code, then move to commit #2, etc. It runs against cassandra so it will demonstrate replaying events that need to be evolved using new schema, all without using an external schema registry.
In my project, I need to ship plenty of existing Java objects to Spark workers, most of them are not extended from java.io.Serializable. I also want the ability to control the variables/attributes included in the objects. I only want to serialize the useful attributes, not everything in a object.
Spark documentation indicates that there are two ways to serialize an object in Spark, using java.io.Serializable or Kryo. I think both ways need to rewrite a bunch of wrappers or extra code for every business objects. However, my current code base has already implemented protocol-buffers serialization. I am wondering if there is any way to embed this serialization mechanism into Spark.