Apache Flink：“类X不包含字段Y的设置器”是什么意思？

Question

我第一次使用flink（1.6，1.7），并使用来自https://www.gharchive.org/的github档案中的数据，但将这些数据用作流数据源。

我的简单示例仅计算每个用户每天的所有事件，而我尝试复制相同的示例，但改用TableEnvironment和SQL支持。

但是，我遇到以下错误：

org.apache.flink.streaming.api.functions.source.TimestampedFileInputSplit类不包含用于字段ModifyTime的设置器，如下所示：

 8-12-04 14:17:02:115  INFO main exploration.StreamingTableApp:32 - Starting Streaming Table Flink App Example...
18-12-04 14:17:02:174  INFO main typeutils.TypeExtractor:1818 - class org.apache.flink.streaming.api.functions.source.TimestampedFileInputSplit does not contain a setter for field modificationTime
18-12-04 14:17:02:176  INFO main typeutils.TypeExtractor:1857 - Class class org.apache.flink.streaming.api.functions.source.TimestampedFileInputSplit cannot be used as a POJO type because not all fields are valid POJO fields, and must be processed as GenericType. Please read the Flink documentation on "Data Types & Serialization" for details of the effect on performance.
18-12-04 14:17:02:937  INFO main exploration.StreamingTableApp:74 - Finished...

我正在读取CSV源作为数据流，并使用Gson解析出json行的位并将这些属性映射到元组。

有人对此有任何想法/经验吗？

主要方法：

StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();

env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime);

// Mapped in docker compose file too.
DataStreamSource<String> input = env.readTextFile("/some/path/github/");

// Setup the stream
DataStream<Tuple4<String, Integer, String, Long>> stream = input.map(new GithubTupleConverter())
    .assignTimestampsAndWatermarks(new TupleTimestampExtractor());

StreamTableEnvironment tEnv = TableEnvironment.getTableEnvironment(env);
Table tableFromStream = tEnv.fromDataStream(stream, "user_id, kount, basic_date,event_date");

TupleTimestampExtractor

public class TupleTimestampExtractor
        extends BoundedOutOfOrdernessTimestampExtractor<Tuple4<String, Integer, String, Long>> {
    private static final long serialVersionUID = 3737675541965835242L;

    public TupleTimestampExtractor() {
        super(Time.seconds(30L));
    }

    @Override
    public long extractTimestamp(Tuple4<String, Integer, String, Long> element) {
        return element.getField(3);
    }
}

GithubTupleConverter.java

public class GithubTupleConverter implements MapFunction<String, Tuple4<String, Integer, String, Long>> {

    private static final Gson g = new Gson();

    @Override
    public Tuple4<String, Integer, String, Long> map(String value) throws Exception {
        // Take each line as Json.
        JsonObject o = g.fromJson(value, JsonObject.class);

        // Extract the user id
        String userId = o.get("actor").getAsJsonObject().get("login").getAsString();

        // Extract the event type (commit, pull request, fork event)
        String type = o.get("type").getAsString();

        // Get the event date time
        String dateTime = o.get("created_at").getAsString();

        // Parse date string to Typed type.
        LocalDateTime eventTime = LocalDateTime.parse(dateTime, DateTimeFormatter.ISO_DATE_TIME);

        // Format the date so it can be used in the output.
        DateTimeFormatter formatter = DateTimeFormatter.ISO_DATE;

        return Tuple4.of(userId, 1, formatter.format(eventTime), eventTime.toInstant(ZoneOffset.UTC).toEpochMilli());
    }
} 

Answer 1

您共享的日志不会显示错误。日志处于INFO级别，不会引发异常（至少不在提供的日志中）。

日志条目仅表明类TimestampedFileInputSplit不能被视为POJO。通常，此消息表明性能不是最佳的，但是在这种特定情况下，这不是问题。

您还收到其他任何错误消息吗？