Spring Kafka过滤器不过滤消费者记录

Question

我正在尝试根据 ConsumerRecord 中某个字段的内容在消费前过滤 ConsumerRecord 消息。

应用过滤器之前的示例消费者记录是（在值中查找 GP_ID）：

 ConsumerRecord(topic = jdbc-project, partition = 0, offset = 0, CreateTime = 1551118248440, serialized key size = -1, serialized value size = 69, headers = RecordHeaders(headers = [], isReadOnly = false), key = null, value = {"GP_ID": {"bytes": "@"}, "PROJECT_ID": {"bytes": "\u001E\u008C"}, "START_DATE": 1009843200000, "END_DATE": 1041292800000, "TITLE": "Project- FPH", "STATUS_CODE": "INACTIVE"})

KafkaRecordVO(projectId=7820, gpId=64)

当我在 kafkaListenerContainerFactory() 中设置如下记录过滤策略时：

@Bean
ConcurrentKafkaListenerContainerFactory<String, GenericRecord> kafkaListenerContainerFactoryProject() {
    ConcurrentKafkaListenerContainerFactory<String, GenericRecord> factory = new ConcurrentKafkaListenerContainerFactory<>();
    factory.setConsumerFactory(consumerFactory());
    factory.setRecordFilterStrategy(new RecordFilterStrategy<String, GenericRecord>() {
        @Override
        public boolean filter(ConsumerRecord<String, GenericRecord> consumerRecord) {
            long gpId= KafkaRecordVO.convertByteBufferToLong(consumerRecord.value().get("GP_ID"));
            if(gpId == 10766L || gpId == 10823L || gpId == 10459L || gpId == 10649L)
                return false;
            else
                return true;
        }
    });
    factory.getContainerProperties().setAckMode(AckMode.MANUAL_IMMEDIATE);
    return factory;
}

KafkaRecordVO.convertByteBufferToLong 正在将 bytebuffer 值转换为 long 值。

这是正确评估，并返回真/假。

但是，当它被 Kafka 监听器消费时，如下所示：

@KafkaListener(id = "project", topics = "jdbc-project", containerFactory = "kafkaListenerContainerFactoryProject")
public void consumeProject(ConsumerRecord<String, GenericRecord> record,Acknowledgment acknowledgment) {
    log.debug(record.toString());
    KafkaRecordVO recordVo = new KafkaRecordVO().projectId(record.value().get("PROJECT_ID"))
                                                .budgetYear(record.value().get("GP_ID"));
    log.debug(recordVo.toString());
}

这将返回删除我过滤的字段值的记录：“GP_ID”

这些是应用过滤器后生成的示例日志（在值中查找 GP_ID）：

ConsumerRecord(topic = jdbc-project, partition = 0, offset = 171275, CreateTime = 1551118279371, serialized key size = -1, serialized value size = 181, headers = RecordHeaders(headers = [], isReadOnly = false), key = null, value = { "GP_ID": {"bytes": ""}, "PROJECT_ID": {"bytes": "\u0005â^"}, "START_DATE": 1470009600000, "END_DATE": 1532995200000, "TITLE": "Project 2016 - 2016", "STATUS_CODE": "INACTIVE"})

KafkaRecordVO(projectId=385630, gpId=0)

我在我的 kafkaListener 中为这个主题在其他字段的日志中得到了这个：“GP_ID”：{“bytes”：“”}。如何不剥离价值？这里有什么问题？

更新：问题发生在我使用实用方法将字节转换为长整数时。它将位置设置为数组的末尾。这就是它返回空数组的原因：

public static Long convertByteBufferToLong(Object byteBuff) {
    //After adding below line, the issue got resolved
    ByteBuffer buf = ((ByteBuffer) byteBuff).duplicate();
    byte[] arr = new byte[buf.remaining()];
    buf.get(arr);
    BigInteger bi =new BigInteger(1,arr);
    return bi.longValue();
}

Answer 1

你的建议毫无意义；过滤器适配器有这个代码...

@Override
public void onMessage(ConsumerRecord<K, V> consumerRecord, Acknowledgment acknowledgment, Consumer<?, ?> consumer) {
    if (!filter(consumerRecord)) {
        switch (this.delegateType) {
            case ACKNOWLEDGING_CONSUMER_AWARE:
                this.delegate.onMessage(consumerRecord, acknowledgment, consumer);
                break;
            case ACKNOWLEDGING:
                this.delegate.onMessage(consumerRecord, acknowledgment);
                break;
            case CONSUMER_AWARE:
                this.delegate.onMessage(consumerRecord, consumer);
                break;
            case SIMPLE:
                this.delegate.onMessage(consumerRecord);
        }
    }
    else {
        ackFilteredIfNecessary(acknowledgment);
    }
}

它不对记录进行任何操作。

尝试在调试器中运行，看看是否可以看到发生了什么。