java - Apache Beam 和 CombineFn 的编​​码器问题

Question

我们正在使用 Apache Beam 和 DirectRunner 构建管道作为运行者。我们目前正在尝试一个简单的管道，我们可以:

1. 从 Google Cloud Pub/Sub 中提取数据(目前使用模拟器在本地运行)
2. 反序列化为 Java 对象
3. 使用 1 分钟固定窗口的窗口事件
4. 使用自定义 CombineFn 组合这些窗口将它们从事件转换为事件列表。

---

管道代码:

pipeline
.apply(PubsubIO.<String>read().topic(options.getTopic()).withCoder(StringUtf8Coder.of()))

.apply("ParseEvent", ParDo.of(new ParseEventFn()))

.apply("WindowOneMinute",Window.<Event>into(FixedWindows.of(Duration.standardMinutes(1))))              

.apply("CombineEvents", Combine.globally(new CombineEventsFn()));

---

ParseEvent 函数:

    static class ParseEventFn extends DoFn<String, Event> {
        @ProcessElement
        public void processElement(ProcessContext c) {
            String json = c.element();
            c.output(gson.fromJson(json, Event.class));
        }
    }

---

CombineEvents 函数:

public static class CombineEventsFn extends CombineFn<Event, CombineEventsFn.Accum, EventListWrapper> {
        public static class Accum {
            EventListWrapper eventListWrapper = new EventListWrapper();
        }

        @Override
        public Accum createAccumulator() {
            return new Accum();
        }

        @Override
        public Accum addInput(Accum accumulator, Event event) {
            accumulator.eventListWrapper.events.add(event);
            return accumulator;
        }

        @Override
        public Accum mergeAccumulators(Iterable<Accum> accumulators) {
            Accum merged = createAccumulator();
            for (Accum accum : accumulators) {
                merged.eventListWrapper.events.addAll(accum.eventListWrapper.events);
            }
            return merged;
        }

        @Override
        public EventListWrapper extractOutput(Accum accumulator) {
            return accumulator.eventListWrapper;
        }

    }

---

尝试使用 Maven 和 DirectRunner 在本地运行时，我们收到以下错误:

java.lang.reflect.InvocationTargetException
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at org.codehaus.mojo.exec.ExecJavaMojo$1.run(ExecJavaMojo.java:293)
    at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.IllegalStateException: Unable to return a default Coder for CombineEvents/Combine.perKey(CombineEvents)/Combine.GroupedValues/ParDo(Anonymous).out [PCollection]. Correct one of the following root causes:
  No Coder has been manually specified;  you may do so using .setCoder().
  Inferring a Coder from the CoderRegistry failed: Unable to provide a default Coder for org.apache.beam.sdk.values.KV<K, OutputT>. Correct one of the following root causes:
  Building a Coder using a registered CoderFactory failed: Cannot provide coder for parameterized type org.apache.beam.sdk.values.KV<K, OutputT>: Unable to provide a default Coder for java.lang.Object. Correct one of the following root causes:
  Building a Coder using a registered CoderFactory failed: Cannot provide coder based on value with class java.lang.Object: No CoderFactory has been registered for the class.
  Building a Coder from the @DefaultCoder annotation failed: Class java.lang.Object does not have a @DefaultCoder annotation.
  Building a Coder from the fallback CoderProvider failed: Cannot provide coder for type java.lang.Object: org.apache.beam.sdk.coders.protobuf.ProtoCoder$2@6e610150 could not provide a Coder for type java.lang.Object: Cannot provide ProtoCoder because java.lang.Object is not a subclass of com.google.protobuf.Message; org.apache.beam.sdk.coders.SerializableCoder$1@7adc59c8 could not provide a Coder for type java.lang.Object: Cannot provide SerializableCoder because java.lang.Object does not implement Serializable.
  Building a Coder from the @DefaultCoder annotation failed: Class org.apache.beam.sdk.values.KV does not have a @DefaultCoder annotation.
  Using the default output Coder from the producing PTransform failed: Unable to provide a default Coder for org.apache.beam.sdk.values.KV<K, OutputT>. Correct one of the following root causes:
  Building a Coder using a registered CoderFactory failed: Cannot provide coder for parameterized type org.apache.beam.sdk.values.KV<K, OutputT>: Unable to provide a default Coder for java.lang.Object. Correct one of the following root causes:
  Building a Coder using a registered CoderFactory failed: Cannot provide coder based on value with class java.lang.Object: No CoderFactory has been registered for the class.
  Building a Coder from the @DefaultCoder annotation failed: Class java.lang.Object does not have a @DefaultCoder annotation.
  Building a Coder from the fallback CoderProvider failed: Cannot provide coder for type java.lang.Object: org.apache.beam.sdk.coders.protobuf.ProtoCoder$2@6e610150 could not provide a Coder for type java.lang.Object: Cannot provide ProtoCoder because java.lang.Object is not a subclass of com.google.protobuf.Message; org.apache.beam.sdk.coders.SerializableCoder$1@7adc59c8 could not provide a Coder for type java.lang.Object: Cannot provide SerializableCoder because java.lang.Object does not implement Serializable.
  Building a Coder from the @DefaultCoder annotation failed: Class org.apache.beam.sdk.values.KV does not have a @DefaultCoder annotation.
    at org.apache.beam.sdk.repackaged.com.google.common.base.Preconditions.checkState(Preconditions.java:444)
    at org.apache.beam.sdk.values.TypedPValue.getCoder(TypedPValue.java:51)
    at org.apache.beam.sdk.values.PCollection.getCoder(PCollection.java:130)
    at org.apache.beam.sdk.values.TypedPValue.finishSpecifying(TypedPValue.java:90)
    at org.apache.beam.sdk.runners.TransformHierarchy.finishSpecifyingInput(TransformHierarchy.java:143)
    at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:418)
    at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:334)
    at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:154)
    at org.apache.beam.sdk.transforms.Combine$Globally.expand(Combine.java:1459)
    at org.apache.beam.sdk.transforms.Combine$Globally.expand(Combine.java:1336)
    at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:420)
    at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:350)
    at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:167)
at ***************************.main(***************.java:231)
... 6 more

---

为巨大的代码转储道歉 - 想要提供所有上下文。

我很好奇为什么它提示没有默认编码器 java.lang.Object和 org.apache.beam.sdk.values.KV<K, OutputT> - 据我所知，我们的管道正在 String 之间改变类型, Event , 和 EventListWrapper - 后两个类在类本身上设置了它们的默认编码器(在这两种情况下都是 AvroCoder)。

错误发生在我们应用 CombineFn 的那一行 - 可以确认没有这个转换管道工作。

我怀疑我们以某种方式错误地设置了组合转换，但到目前为止，在 Beam 文档中没有发现任何东西可以为我们指明正确的方向。

如有任何见解，我们将不胜感激 - 提前致谢!

Answer 1

您看到 java.lang.Object 的可能原因是 Beam 正在尝试为未解析的类型变量推断编码器，该变量将被解析为 Object。这可能是编码器如何在 Combine 中进行推理的错误。

另外，我预计 Accum 类也会导致编码器推理失败。您可以覆盖 CombineFn 中的 getAccumulatorCoder 以非常直接地提供一个。