首先，我们故意制造一个异常

然后，发送一个JSON数据，并且缺失了一些字段，看看报什么错！

顺利抓到调用栈

[2018-11-23 13:24:32,877] INFO Source: MyKafka010JsonTableSource -> from: (pro, throwable, level, ip, SPT) -> where: (AND(IS NOT NULL(pro), IS NOT NULL(throwable), IS NOT NULL(level), IS NOT NULL(ip))), select: (pro, throwable, level, ip, SPT) (3/3) (7da5fcb7cd77ff35b64d8bbe367a84d7) switched from RUNNING to FAILED. org.apache.flink.runtime.taskmanager.Task.transitionState(Task.java:927) 
java.io.IOException: Failed to deserialize JSON object.
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:97)
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:50)
	at org.apache.flink.streaming.util.serialization.KeyedDeserializationSchemaWrapper.deserialize(KeyedDeserializationSchemaWrapper.java:44)
	at org.apache.flink.streaming.connectors.kafka.internal.Kafka09Fetcher.runFetchLoop(Kafka09Fetcher.java:142)
	at org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumerBase.run(FlinkKafkaConsumerBase.java:721)
	at org.apache.flink.streaming.api.operators.StreamSource.run(StreamSource.java:87)
	at org.apache.flink.streaming.api.operators.StreamSource.run(StreamSource.java:56)
	at org.apache.flink.streaming.runtime.tasks.SourceStreamTask.run(SourceStreamTask.java:99)
	at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:306)
	at org.apache.flink.runtime.taskmanager.Task.run(Task.java:703)
	at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.IllegalStateException: Could not find field with name 'ip'.
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.convertRow(JsonRowDeserializationSchema.java:189)
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:95)
	... 10 more
[2018-11-23 13:24:32,877] INFO Source: MyKafka010JsonTableSource -> from: (pro, throwable, level, ip, SPT) -> where: (AND(IS NOT NULL(pro), IS NOT NULL(throwable), IS NOT NULL(level), IS NOT NULL(ip))), select: (pro, throwable, level, ip, SPT) (2/3) (f68577aede6953de859c636d9d3b6f3e) switched from RUNNING to FAILED. org.apache.flink.runtime.taskmanager.Task.transitionState(Task.java:927) 
java.io.IOException: Failed to deserialize JSON object.
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:97)
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:50)
	at org.apache.flink.streaming.util.serialization.KeyedDeserializationSchemaWrapper.deserialize(KeyedDeserializationSchemaWrapper.java:44)
	at org.apache.flink.streaming.connectors.kafka.internal.Kafka09Fetcher.runFetchLoop(Kafka09Fetcher.java:142)
	at org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumerBase.run(FlinkKafkaConsumerBase.java:721)
	at org.apache.flink.streaming.api.operators.StreamSource.run(StreamSource.java:87)
	at org.apache.flink.streaming.api.operators.StreamSource.run(StreamSource.java:56)
	at org.apache.flink.streaming.runtime.tasks.SourceStreamTask.run(SourceStreamTask.java:99)
	at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:306)
	at org.apache.flink.runtime.taskmanager.Task.run(Task.java:703)
	at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.IllegalStateException: Could not find field with name 'ip'.
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.convertRow(JsonRowDeserializationSchema.java:189)
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:95)
	... 10 more
[2018-11-23 13:24:32,877] INFO Source: MyKafka010JsonTableSource -> from: (pro, throwable, level, ip, SPT) -> where: (AND(IS NOT NULL(pro), IS NOT NULL(throwable), IS NOT NULL(level), IS NOT NULL(ip))), select: (pro, throwable, level, ip, SPT) (1/3) (8a2c89aa3a2ea8f4c4423c69b22b2522) switched from RUNNING to FAILED. org.apache.flink.runtime.taskmanager.Task.transitionState(Task.java:927) 
java.io.IOException: Failed to deserialize JSON object.
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:97)
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:50)
	at org.apache.flink.streaming.util.serialization.KeyedDeserializationSchemaWrapper.deserialize(KeyedDeserializationSchemaWrapper.java:44)
	at org.apache.flink.streaming.connectors.kafka.internal.Kafka09Fetcher.runFetchLoop(Kafka09Fetcher.java:142)
	at org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumerBase.run(FlinkKafkaConsumerBase.java:721)
	at org.apache.flink.streaming.api.operators.StreamSource.run(StreamSource.java:87)
	at org.apache.flink.streaming.api.operators.StreamSource.run(StreamSource.java:56)
	at org.apache.flink.streaming.runtime.tasks.SourceStreamTask.run(SourceStreamTask.java:99)
	at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:306)
	at org.apache.flink.runtime.taskmanager.Task.run(Task.java:703)
	at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.IllegalStateException: Could not find field with name 'ip'.
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.convertRow(JsonRowDeserializationSchema.java:189)
	at org.apache.flink.formats.json.JsonRowDeserializationSchema.deserialize(JsonRowDeserializationSchema.java:95)
	... 10 more
[2018-11-23 13:24:32,892] INFO Freeing task resources for Source: MyKafka010JsonTableSource -> from: (pro, throwable, level, ip, SPT) -> where: (AND(IS NOT NULL(pro), IS NOT NULL(throwable), IS NOT NULL(level), IS NOT NULL(ip))), select: (pro, throwable, level, ip, SPT) (1/3) (8a2c89aa3a2ea8f4c4423c69b22b2522). org.apache.flink.runtime.taskmanager.Task.run(Task.java:810) 

这里只是为了让大家对内部序列化的地点有个印象。

如果你看懂了，下面进入正题

1）自定义的TableSource中改写

protected FlinkKafkaConsumerBase<Row> createKafkaConsumer(String topic, Properties properties,
                                                              DeserializationSchema<Row> deserializationSchema) {
        //下面2行本质是一样的,但是因为需要自定义,所以自己返回
        //super.createKafkaConsumer(topic, properties, deserializationSchema);  
        return new FlinkKafkaConsumer010<>(Collections.singletonList(topic),
            new MyKeyedDeserializationSchemaWrapper<>(deserializationSchema), properties);
    }

2）MyKeyedDeserializationSchemaWrapper.java

package org.apache.flink.api.java.io.kafka;

import java.io.IOException;

import org.apache.flink.api.common.serialization.DeserializationSchema;
import org.apache.flink.api.java.typeutils.RowTypeInfo;
import org.apache.flink.streaming.util.serialization.KeyedDeserializationSchemaWrapper;
import org.apache.flink.types.Row;

public class MyKeyedDeserializationSchemaWrapper<T> extends KeyedDeserializationSchemaWrapper<T> {

    //    private static final org.slf4j.Logger Logger                = LoggerFactory
    //        .getLogger(MyKeyedDeserializationSchemaWrapper.class);

    private static final int    NOT_FOUND             = -1;
    /**  */
    private static final long   serialVersionUID      = -6984024233082983820L;
    //
    private static final String KAFKA_TOPIC           = "KAFKA_TOPIC";
    private int                 KAFKA_TOPIC_INDEX     = NOT_FOUND;
    //
    private static final String KAFKA_PARTITION       = "KAFKA_PARTITION";
    private int                 KAFKA_PARTITION_INDEX = NOT_FOUND;
    //
    private static final String KAFKA_OFFSET          = "KAFKA_OFFSET";
    private int                 KAFKA_OFFSET_INDEX    = NOT_FOUND;

    //
    private boolean             inited                = false;

    public MyKeyedDeserializationSchemaWrapper(DeserializationSchema<T> deserializationSchema) {
        super(deserializationSchema);
    }

    @SuppressWarnings("unchecked")
    @Override
    public T deserialize(byte[] messageKey, byte[] message, String topic, int partition,
                         long offset) throws IOException {
        Row row = (Row) super.deserialize(messageKey, message, topic, partition, offset);
        //只执行1次
        if (false == inited) {
            RowTypeInfo rowTypeInfo = (RowTypeInfo) super.getProducedType();//Row(pro: String, throwable: String, level: String, ip: String, SPT: Timestamp)
            KAFKA_TOPIC_INDEX = rowTypeInfo.getFieldIndex(KAFKA_TOPIC);
            KAFKA_PARTITION_INDEX = rowTypeInfo.getFieldIndex(KAFKA_PARTITION);
            KAFKA_OFFSET_INDEX = rowTypeInfo.getFieldIndex(KAFKA_OFFSET);
            inited = true;
        }
        //每次都要执行一遍
        if (NOT_FOUND != KAFKA_TOPIC_INDEX) {
            row.setField(KAFKA_TOPIC_INDEX, topic);
        }
        if (NOT_FOUND != KAFKA_PARTITION_INDEX) {
            row.setField(KAFKA_PARTITION_INDEX, partition);
        }
        if (NOT_FOUND != KAFKA_OFFSET_INDEX) {
            row.setField(KAFKA_OFFSET_INDEX, offset);
        }
        //返回最终结果
        return (T) row;
    }

}

测试一把看看效果！

完美！