3

我想加入两个 PCollection（分别来自不同的输入）并按照此处描述的步骤实现，“加入 CoGroupByKey”部分： https ://cloud.google.com/dataflow/model/group-by-key

就我而言，我想加入 GeoIP 的“区块”信息和“位置”信息。因此，我将 Block 和 Location 定义为自定义类，然后编写如下：

final TupleTag<Block> t1 = new TupleTag<Block>();
final TupleTag<Location> t2 = new TupleTag<Location>();
PCollection<KV<Long, CoGbkResult>> coGbkResultColl = KeyedPCollectionTuple.of(t1, kvGeoNameIDBlock)
        .and(t2, kvGeoNameIDLocation).apply(CoGroupByKey.<Long>create());

键具有 Long 类型的值。我以为它已经完成了，但是当我运行时mvn compile，它会输出以下错误：

[ERROR] Failed to execute goal org.codehaus.mojo:exec-maven-plugin:1.4.0:java (default-cli) on project xxxx: An exception occured while executing the Java class. null: InvocationTargetException: Unable to return a default Coder for Extract GeoNameID-Block KV/ParMultiDo(ExtractGeoNameIDBlock).out0 [PCollection]. Correct one of the following root causes:
[ERROR]   No Coder has been manually specified;  you may do so using .setCoder().
[ERROR]   Inferring a Coder from the CoderRegistry failed: Cannot provide coder for parameterized type org.apache.beam.sdk.values.KV<java.lang.Long, com.xxx.platform.geoip2.Block>: Unable to provide a Coder for com.xxx.platform.geoip2.Block.
[ERROR]   Building a Coder using a registered CoderProvider failed.
[ERROR]   See suppressed exceptions for detailed failures.
[ERROR]   Using the default output Coder from the producing PTransform failed: Cannot provide coder for parameterized type org.apache.beam.sdk.values.KV<java.lang.Long, com.xxx.platform.geoip2.Block>: Unable to provide a Coder for com.xxx.platform.geoip2.Block.

输出错误的确切 DoFn 是ExtractGeoNameIDBlock，它只是创建其键（要连接）和自身的键值对。

// ExtractGeoNameIDBlock creates KV collection while reading from block CSV
static class ExtractGeoNameIDBlock extends DoFn<String, KV<Long, Block>> {
private static final long serialVersionUID = 1L;

  @ProcessElement
  public void processElement(ProcessContext c) throws Exception {
    String line = c.element();

    if (!line.startsWith("network,")) { // exclude headerline
      Block b = new Block();
      b.loadFromCsvLine(line);

      if (b.getGeonameId() != null) {
        c.output(KV.of(b.getGeonameId(), b));
      }
    }
  }
}

loadFromCsvLine只需解析 CSV 行，将字段转换为每个相应的类型并分配给它的私有字段。

所以看起来我需要为我的自定义类设置一些编码器才能使其工作。我找到了一个引用编码器的文档，但仍然不确定如何实现我的。 https://cloud.google.com/dataflow/model/data-encoding

有没有我可以遵循的真实示例来为我的自定义类创建自定义编码器？

[更新 13:02 09/26/2017] 我添加了

CoderRegistry cr = p.getCoderRegistry();
cr.registerCoderForClass(Block.class, AvroCoder.of(Block.class));

然后得到一个错误

 java.lang.NullPointerException: in com.xxx.platform.geoip2.Block in long null of long in field representedCountryGeonameId of com.xxx.platform.geoip2.Block

[更新 14:05 09/26/2017] 我改变了这样的实现：

@DefaultCoder(AvroCoder.class)
public class Block {
    private static final Logger LOG = LoggerFactory.getLogger(Block.class);

    @Nullable
    public String network;
    @Nullable
    public Long registeredCountryGeonameId;
:
:

（将@Nullable 设置为所有属性）

但仍然出现此错误：

(22eeaf3dfb26f8cc): java.lang.RuntimeException: org.apache.beam.sdk.coders.CoderException: cannot encode a null Long
    at com.google.cloud.dataflow.worker.SimpleParDoFn$1.output(SimpleParDoFn.java:191)
    at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
    at org.apache.beam.runners.core.SimpleDoFnRunner.access$700(SimpleDoFnRunner.java:66)
    at org.apache.beam.runners.core.SimpleDoFnRunner$DoFnProcessContext.output(SimpleDoFnRunner.java:436)
    at org.apache.beam.runners.core.SimpleDoFnRunner$DoFnProcessContext.output(SimpleDoFnRunner.java:424)
    at org.apache.beam.sdk.transforms.join.CoGroupByKey$ConstructUnionTableFn.processElement(CoGroupByKey.java:185)
Caused by: org.apache.beam.sdk.coders.CoderException: cannot encode a null Long
    at org.apache.beam.sdk.coders.VarLongCoder.encode(VarLongCoder.java:51)
    at org.apache.beam.sdk.coders.VarLongCoder.encode(VarLongCoder.java:35)
    at org.apache.beam.sdk.coders.Coder.encode(Coder.java:135)
    at com.google.cloud.dataflow.worker.ShuffleSink$ShuffleSinkWriter.encodeToChunk(ShuffleSink.java:320)
    at com.google.cloud.dataflow.worker.ShuffleSink$ShuffleSinkWriter.add(ShuffleSink.java:216)
    at com.google.cloud.dataflow.worker.ShuffleSink$ShuffleSinkWriter.add(ShuffleSink.java:178)
    at com.google.cloud.dataflow.worker.util.common.worker.WriteOperation.process(WriteOperation.java:80)
    at com.google.cloud.dataflow.worker.util.common.worker.OutputReceiver.process(OutputReceiver.java:52)
    at com.google.cloud.dataflow.worker.ReifyTimestampAndWindowsParDoFnFactory$ReifyTimestampAndWindowsParDoFn.processElement(ReifyTimestampAndWindowsParDoFnFactory.java:68)
    at com.google.cloud.dataflow.worker.util.common.worker.ParDoOperation.process(ParDoOperation.java:48)
    at com.google.cloud.dataflow.worker.util.common.worker.OutputReceiver.process(OutputReceiver.java:52)
    at com.google.cloud.dataflow.worker.SimpleParDoFn$1.output(SimpleParDoFn.java:183)
    at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
    at org.apache.beam.runners.core.SimpleDoFnRunner.access$700(SimpleDoFnRunner.java:66)
    at org.apache.beam.runners.core.SimpleDoFnRunner$DoFnProcessContext.output(SimpleDoFnRunner.java:436)
    at org.apache.beam.runners.core.SimpleDoFnRunner$DoFnProcessContext.output(SimpleDoFnRunner.java:424)
    at org.apache.beam.sdk.transforms.join.CoGroupByKey$ConstructUnionTableFn.processElement(CoGroupByKey.java:185)
    at org.apache.beam.sdk.transforms.join.CoGroupByKey$ConstructUnionTableFn$DoFnInvoker.invokeProcessElement(Unknown Source)
    at org.apache.beam.runners.core.SimpleDoFnRunner.invokeProcessElement(SimpleDoFnRunner.java:177)
    at org.apache.beam.runners.core.SimpleDoFnRunner.processElement(SimpleDoFnRunner.java:141)
    at com.google.cloud.dataflow.worker.SimpleParDoFn.processElement(SimpleParDoFn.java:233)
    at com.google.cloud.dataflow.worker.util.common.worker.ParDoOperation.process(ParDoOperation.java:48)
    at com.google.cloud.dataflow.worker.util.common.worker.OutputReceiver.process(OutputReceiver.java:52)
    at com.google.cloud.dataflow.worker.SimpleParDoFn$1.output(SimpleParDoFn.java:183)
    at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
    at org.apache.beam.runners.core.SimpleDoFnRunner.access$700(SimpleDoFnRunner.java:66)
    at org.apache.beam.runners.core.SimpleDoFnRunner$DoFnProcessContext.output(SimpleDoFnRunner.java:436)
    at org.apache.beam.runners.core.SimpleDoFnRunner$DoFnProcessContext.output(SimpleDoFnRunner.java:424)
    at com.bandainamcoent.platform.GeoIpPopulateTable$ExtractGeoNameIDBlock.processElement(GeoIpPopulateTable.java:79)
    at com.bandainamcoent.platform.GeoIpPopulateTable$ExtractGeoNameIDBlock$DoFnInvoker.invokeProcessElement(Unknown Source)
    at org.apache.beam.runners.core.SimpleDoFnRunner.invokeProcessElement(SimpleDoFnRunner.java:177)
    at org.apache.beam.runners.core.SimpleDoFnRunner.processElement(SimpleDoFnRunner.java:141)
    at com.google.cloud.dataflow.worker.SimpleParDoFn.processElement(SimpleParDoFn.java:233)
    at com.google.cloud.dataflow.worker.util.common.worker.ParDoOperation.process(ParDoOperation.java:48)
    at com.google.cloud.dataflow.worker.util.common.worker.OutputReceiver.process(OutputReceiver.java:52)
    at com.google.cloud.dataflow.worker.util.common.worker.ReadOperation.runReadLoop(ReadOperation.java:187)
    at com.google.cloud.dataflow.worker.util.common.worker.ReadOperation.start(ReadOperation.java:148)
    at com.google.cloud.dataflow.worker.util.common.worker.MapTaskExecutor.execute(MapTaskExecutor.java:68)
    at com.google.cloud.dataflow.worker.DataflowWorker.executeWork(DataflowWorker.java:336)
    at com.google.cloud.dataflow.worker.DataflowWorker.doWork(DataflowWorker.java:294)
    at com.google.cloud.dataflow.worker.DataflowWorker.getAndPerformWork(DataflowWorker.java:244)
    at com.google.cloud.dataflow.worker.DataflowBatchWorkerHarness$WorkerThread.doWork(DataflowBatchWorkerHarness.java:135)
    at com.google.cloud.dataflow.worker.DataflowBatchWorkerHarness$WorkerThread.call(DataflowBatchWorkerHarness.java:115)
    at com.google.cloud.dataflow.worker.DataflowBatchWorkerHarness$WorkerThread.call(DataflowBatchWorkerHarness.java:102)
    at java.util.concurrent.FutureTask.run(FutureTask.java:266)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:745)

谢谢。

4

2 回答 2

1

看起来您的自定义类Block没有指定编码器。您可以创建自己的Coder，也可以使用其中一种通用的，例如AvroCoder. 您还应该使用 s 注册它，CoderRegistry以便管道知道如何对Blocks 进行编码。

于 2017-09-26T19:01:39.650 回答
0

我终于在我的问题中于 2017 年 9 月 26 日 14:05 更新时使用 AvroCoder + Nullable 注释做到了这一点。

我看到的最后一个错误只是因为我的数据实际上有一个我没想到的空值。在我的 Java 代码中处理空值后，一切正常。

我认为关于另一个问题的这篇文章对这个问题非常有用： https ://stackoverflow.com/a/32342403/2543803

于 2017-09-26T23:39:18.830 回答