Kafka stream consumer for JSON object : How to map

Question

I am new to Kafka/Kafka Stream. I am using latest Kafka/kafka-stream and kafka-client and openjdk11. My producer is producing json objects ( Where key is the name) that looks like

{"Name":"John", "amount":123, "time":2019-10-03T05:24:52" }

Producer code for better understanding:

public static ProducerRecord<String, String> newRandomTransaction(String name) {
    // creates an empty json {}
    ObjectNode transaction = JsonNodeFactory.instance.objectNode();

    Integer amount = ThreadLocalRandom.current().nextInt(0, 100);

    // Instant.now() is to get the current time
    Instant now = Instant.now();

    // we write the data to the json document
    transaction.put("name", name);
    transaction.put("amount", amount);
    transaction.put("time", now.toString());
    return new ProducerRecord<>("bank-transactions", name, transaction.toString());
}

Now I am trying to write my application that consumes the transactions and compute the total money in that person's balance.

( FYI: I am using an old code and trying to make it work).

Used GroupBYKey as the topic already has the right key. And then aggregate to compute the total balance where I am struggling.

Application at this moment ( commented out part is the old code that I am trying to make it work in the next line):

public class BankBalanceExactlyOnceApp {
    private static ObjectMapper mapper = new ObjectMapper();

    public static void main(String[] args) {
        Properties config = new Properties();

        config.put(StreamsConfig.APPLICATION_ID_CONFIG, "bank-balance-application");
        config.put(StreamsConfig.BOOTSTRAP_SERVERS_CONFIG, "127.0.0.1:9092");
        config.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "earliest");

        // we disable the cache to demonstrate all the "steps" involved in the transformation - not recommended in prod
        config.put(StreamsConfig.CACHE_MAX_BYTES_BUFFERING_CONFIG, "0");

        // Exactly once processing!!
        config.put(StreamsConfig.PROCESSING_GUARANTEE_CONFIG, StreamsConfig.EXACTLY_ONCE);

        // json Serde
        final Serializer<JsonNode> jsonSerializer = new JsonSerializer();
        final Deserializer<JsonNode> jsonDeserializer = new JsonDeserializer();
        final Serde<JsonNode> jsonSerde = Serdes.serdeFrom(jsonSerializer, jsonDeserializer);


        StreamsBuilder builder = new StreamsBuilder();

        KStream<String, JsonNode> bankTransactions =
                builder.stream( "bank-transactions", Materialized.with(Serdes.String(), jsonSerde);


        // create the initial json object for balances
        ObjectNode initialBalance = JsonNodeFactory.instance.objectNode();
        initialBalance.put("count", 0);
        initialBalance.put("balance", 0);
        initialBalance.put("time", Instant.ofEpochMilli(0L).toString());

        /*KTable<String, JsonNode> bankBalance = bankTransactions
                .groupByKey(Serdes.String(), jsonSerde)
                .aggregate(
                        () -> initialBalance,
                        (key, transaction, balance) -> newBalance(transaction, balance),
                        jsonSerde,
                        "bank-balance-agg"
                );*/

        KTable<String, JsonNode> bankBalance = bankTransactions
                .groupByKey(Serialized.with(Serdes.String(), jsonSerde))
                .aggregate(
                        () -> initialBalance,
                        (key, transaction, balance) -> {
                            //String t = transaction.toString();
                            newBalance(transaction, balance);
                        },
                        Materialized.with(Serdes.String(), jsonSerde),
                        "bank-balance-agg"
                );

        bankBalance.toStream().to("bank-balance-exactly-once", Produced.with(Serdes.String(), jsonSerde));

        KafkaStreams streams = new KafkaStreams(builder.build(), config);
        streams.cleanUp();
        streams.start();

        // print the topology
        System.out.println(streams.toString());

        // shutdown hook to correctly close the streams application
        Runtime.getRuntime().addShutdownHook(new Thread(streams::close));
    }

    private static JsonNode newBalance(JsonNode transaction, JsonNode balance) {
        // create a new balance json object
        ObjectNode newBalance = JsonNodeFactory.instance.objectNode();
        newBalance.put("count", balance.get("count").asInt() + 1);
        newBalance.put("balance", balance.get("balance").asInt() + transaction.get("amount").asInt());

        Long balanceEpoch = Instant.parse(balance.get("time").asText()).toEpochMilli();
        Long transactionEpoch = Instant.parse(transaction.get("time").asText()).toEpochMilli();
        Instant newBalanceInstant = Instant.ofEpochMilli(Math.max(balanceEpoch, transactionEpoch));
        newBalance.put("time", newBalanceInstant.toString());
        return newBalance;
    }
}

The issue is : as I trying to call newBalance(transaction, balance) in the line:

aggregate(
                    () -> initialBalance,
                    (key, transaction, balance) -> newBalance(transaction, balance),
                    jsonSerde,
                    "bank-balance-agg"
            )

and seeing the compiler error with msg:

newBalance(JsonNode, JsonNode) can not be applied to (<lambda parameter>,<lambda parameter>)

I tried to read it as string, changed the param type from JsonNode to Object. However, could not fix it.

May I get any suggestion on how to fix it?

Answer 1

KGroupedStream in Kafka Streams 2.3 doesn't have method with following signature:

<VR> KTable<K, VR> aggregate(final Initializer<VR> initializer,
                             final Aggregator<? super K, ? super V, VR> aggregator,
                             final Materialized<K, VR, KeyValueStore<Bytes, byte[]>> materialized,
                             String aggregateName);

There are two overloaded method aggregate:

<VR> KTable<K, VR> aggregate(final Initializer<VR> initializer,
                             final Aggregator<? super K, ? super V, VR> aggregator);

<VR> KTable<K, VR> aggregate(final Initializer<VR> initializer,
                             final Aggregator<? super K, ? super V, VR> aggregator,
                             final Materialized<K, VR, KeyValueStore<Bytes, byte[]>> materialized);

You should use second one and your code should look something like:

KTable<String, JsonNode> bankBalance = input
        .groupByKey(Grouped.with(Serdes.String(), jsonSerde))
        .aggregate(
                () -> initialBalance,
                (key, transaction, balance) -> newBalance(transaction, balance),
                Materialized.with(Serdes.String(), jsonSerde)
        );