将 CoGroupByKey 与自定义类型一起使用会导致编码器错误
Using CoGroupByKey with custom type ends up in a Coder error
我想加入两个 PCollection（分别来自不同的输入）并按照此处 "Joins with CoGroupByKey" 部分描述的步骤实施：
https://cloud.google.com/dataflow/model/group-by-key
就我而言，我想加入GeoIP的"block"信息和"location"信息。所以我将 Block 和 Location 定义为自定义 class 然后写如下：
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());
一个键有一个 Long 类型的值。我以为已经完成了，但是当我 运行 mvn compile 时，它输出以下错误：
[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，它只是创建一个键值对（要连接的键）和它自己。
// 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 只是解析CSV行，将字段转换为每个对应的类型并分配给它的私有字段。
看来我需要为我的自定义 class 设置一些编码器才能使其正常工作。
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 class 创建自定义编码器？
[更新 13:02 2017 年 9 月 26 日]
我添加了
CoderRegistry cr = p.getCoderRegistry();
cr.registerCoderForClass(Block.class, AvroCoder.of(Block.class));
然后报错
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 2017 年 9 月 26 日]
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.output(SimpleParDoFn.java:191)
at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
at org.apache.beam.runners.core.SimpleDoFnRunner.access0(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$ReifyTimestampAndWindowsParDoFn.processElement(ReifyTimestampAndWindowsParDoFnFactory.java:68)
at com.google.cloud.dataflow.worker.util.common.worker.ParDoOperation.process(ParDoOperation.java:48)
at com.google.cloud.dataflow.worker.util.common.worker.OutputReceiver.process(OutputReceiver.java:52)
at com.google.cloud.dataflow.worker.SimpleParDoFn.output(SimpleParDoFn.java:183)
at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
at org.apache.beam.runners.core.SimpleDoFnRunner.access0(SimpleDoFnRunner.java:66)
at org.apache.beam.runners.core.SimpleDoFnRunner$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.output(SimpleParDoFn.java:183)
at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
at org.apache.beam.runners.core.SimpleDoFnRunner.access0(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。
您的自定义 class Block 似乎没有指定编码器。您可以创建自己的 Coder，或使用通用的 AvroCoder 之一。您还应该使用 CoderRegistry 注册它，以便管道知道如何编码 Blocks。
我终于在 14:05 09/26/2017 post 更新时使用 AvroCoder + Nullable 注释实现了它
在我的问题中。
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。
我认为这个post在另一个问题上对这个问题很有用：
我想加入两个 PCollection（分别来自不同的输入）并按照此处 "Joins with CoGroupByKey" 部分描述的步骤实施： https://cloud.google.com/dataflow/model/group-by-key
就我而言，我想加入GeoIP的"block"信息和"location"信息。所以我将 Block 和 Location 定义为自定义 class 然后写如下：
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());
一个键有一个 Long 类型的值。我以为已经完成了，但是当我 运行 mvn compile 时，它输出以下错误：
[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，它只是创建一个键值对（要连接的键）和它自己。
// 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 只是解析CSV行，将字段转换为每个对应的类型并分配给它的私有字段。
看来我需要为我的自定义 class 设置一些编码器才能使其正常工作。 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 class 创建自定义编码器？
[更新 13:02 2017 年 9 月 26 日] 我添加了
CoderRegistry cr = p.getCoderRegistry();
cr.registerCoderForClass(Block.class, AvroCoder.of(Block.class));
然后报错
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 2017 年 9 月 26 日] 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.output(SimpleParDoFn.java:191)
at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
at org.apache.beam.runners.core.SimpleDoFnRunner.access0(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.output(SimpleParDoFn.java:183)
at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
at org.apache.beam.runners.core.SimpleDoFnRunner.access0(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$DoFnInvoker.invokeProcessElement(Unknown Source)
at org.apache.beam.runners.core.SimpleDoFnRunner.invokeProcessElement(SimpleDoFnRunner.java:177)
at org.apache.beam.runners.core.SimpleDoFnRunner.processElement(SimpleDoFnRunner.java:141)
at com.google.cloud.dataflow.worker.SimpleParDoFn.processElement(SimpleParDoFn.java:233)
at com.google.cloud.dataflow.worker.util.common.worker.ParDoOperation.process(ParDoOperation.java:48)
at com.google.cloud.dataflow.worker.util.common.worker.OutputReceiver.process(OutputReceiver.java:52)
at com.google.cloud.dataflow.worker.SimpleParDoFn.output(SimpleParDoFn.java:183)
at org.apache.beam.runners.core.SimpleDoFnRunner.outputWindowedValue(SimpleDoFnRunner.java:211)
at org.apache.beam.runners.core.SimpleDoFnRunner.access0(SimpleDoFnRunner.java:66)
at org.apache.beam.runners.core.SimpleDoFnRunner$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。
您的自定义 class Block 似乎没有指定编码器。您可以创建自己的 Coder，或使用通用的 AvroCoder 之一。您还应该使用 CoderRegistry 注册它，以便管道知道如何编码 Blocks。
我终于在 14:05 09/26/2017 post 更新时使用 AvroCoder + Nullable 注释实现了它 在我的问题中。
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。
我认为这个post在另一个问题上对这个问题很有用：