bigquery-kafka-connect

Kafka Connect connector for Google BigQuery

Use API

npm install --save bigquery-kafka-connect

bigquery -> kafka

const { runSourceConnector } = require("bigquery-kafka-connect");
runSourceConnector(config, [], onError).then(config => {
    //runs forever until: config.stop();
});

kafka -> bigquery

const { runSinkConnector } = require("bigquery-kafka-connect");
runSinkConnector(config, [], onError).then(config => {
    //runs forever until: config.stop();
});

kafka -> bigquery (with custom topic (no source-task topic))

const { runSinkConnector, ConverterFactory } = require("bigquery-kafka-connect");
 
const bigQueryTableDescription = {
    "schema": {
        "fields": [
            { name: "id", type: "INTEGER", mode: "REQUIRED" },
            { name: "name", type: "STRING", mode: "REQUIRED" },
            { name: "info", type: "STRING", mode: "NULLABLE" }
        ]
    },
    "timePartitioning": {"type": "DAY"}
};
 
const etlFunc = (messageValue, callback) => {
 
    //type is an example json format field
    if (messageValue.type === "publish") {
        return callback(null, {
            id: messageValue.payload.id,
            name: messageValue.payload.name,
            info: messageValue.payload.info
        });
    }
 
    if (messageValue.type === "unpublish") {
        return callback(null, null); //null value will cause deletion
    }
 
    callback(new Error("unknown messageValue.type"));
};
 
const converter = ConverterFactory.createSinkSchemaConverter(bigQueryTableDescription, etlFunc);
 
runSinkConnector(config, [converter], onError).then(config => {
    //runs forever until: config.stop();
});
 
/*
    this example would be able to store kafka message values
    that look like this (so completely unrelated to messages created by a default SourceTask)
    {
        payload: {
            id: 1,
            name: "first item",
            info: "some info"
        },
        type: "publish"
    }
*/

Use CLI

note: in BETA 🌱

npm install -g bigquery-kafka-connect

# run source etl: bigquery -> kafka
nkc-bigquery-source --help

# run sink etl: kafka -> bigquery
nkc-bigquery-sink --help

Config(uration)

const config = {
    kafka: {
        kafkaHost: "localhost:9092",
        logger: null,
        groupId: "kc-bigquery-test",
        clientName: "kc-bigquery-test-name",
        workerPerPartition: 1,
        options: {
            sessionTimeout: 8000,
            protocol: ["roundrobin"],
            fromOffset: "earliest", //latest
            fetchMaxBytes: 1024 * 100,
            fetchMinBytes: 1,
            fetchMaxWaitMs: 10,
            heartbeatInterval: 250,
            retryMinTimeout: 250,
            requireAcks: 1,
            //ackTimeoutMs: 100,
            //partitionerType: 3
        }
    },
    topic: "sc_test_topic",
    partitions: 1,
    maxTasks: 1,
    pollInterval: 2000,
    produceKeyed: true,
    produceCompressionType: 0,
    connector: {
        batchSize: 500,
        maxPollCount: 500,
        projectId: "bq-project-id",
        dataset: "bq_dataset",
        table: "bq_table",
        idColumn: "id"
    },
    http: {
        port: 3149,
        middlewares: []
    },
    enableMetrics: true,
    batch: {
        batchSize: 100, 
        commitEveryNBatch: 1, 
        concurrency: 1,
        commitSync: true
    }
};

Native Client Config(uration)

const config = {
    kafka: {
        noptions: {
            "metadata.broker.list": "localhost:9092",
            "group.id": "kc-bigquery-test",
            "enable.auto.commit": false,
            "debug": "all",
            "event_cb": true,
            "client.id": "kc-bigquery-test-name"
        },
        tconf: {
            "auto.offset.reset": "earliest",
            "request.required.acks": 1
        }
    },
    topic: "sc_test_topic",
    partitions: 1,
    maxTasks: 1,
    pollInterval: 2000,
    produceKeyed: true,
    produceCompressionType: 0,
    connector: {
        batchSize: 500,
        maxPollCount: 500,
        projectId: "bq-project-id",
        dataset: "bq_dataset",
        table: "bq_table",
        idColumn: "id"
    },
    http: {
        port: 3149,
        middlewares: []
    },
    enableMetrics: true
};