0

我是pyspark的新手，所以希望有人能提供帮助。我正在尝试读取存储在 GCP 存储桶上的镶木地板文件。该文件按日期分区，例如bucket-name/year={}/month={}/day={}

对于给定的文件，我们有以下架构描述：

  1. 直到 3 月，我们以前的 x 和 y 列都是浮点数据类型
  2. 自 3 月以来，这 2 列现在是双数据类型

据我所知，pyspark 在评估float 和 double 数据类型是兼容的数据类型方面没有问题。 （我在网上找到的关于此错误的类似示例与数据类型不兼容有关，例如 String 和 float） 但是，如果我们尝试读取此文件的所有可用数据，我们将面临这个奇怪的问题：

#i.e. read all the data we have ever received for this file
 path = 'bucket-name/year=*/month=*/day=*' 

df = spark.read.format('parquet').load(path)
df.cache().count()

我们得到以下错误。（请注意，如果我们这样做df.count()，我们不会收到此错误，只有在我们先缓存时才会遇到）

此外，从 spark.read 生成的架构将 x 列的数据类型称为浮点数。所以在模式方面，spark 很乐意读入数据并说 dtype 是浮点数。但是，如果我们缓存，事情就会变糟。

希望情况的细节足够清楚:)

An error occurred while calling o923.count. :
org.apache.spark.SparkException: Job aborted due to stage failure:
Task 15 in stage 41.0 failed 4 times, most recent failure: Lost task
15.3 in stage 41.0 (TID 13228, avroconversion-validation-w-1.c.vf-gned-nwp-live.internal, executor
47): java.lang.UnsupportedOperationException:
org.apache.parquet.column.values.dictionary.PlainValuesDictionary$PlainDoubleDictionary
    at
org.apache.parquet.column.Dictionary.decodeToFloat(Dictionary.java:53)
    at
org.apache.spark.sql.execution.datasources.parquet.ParquetDictionary.decodeToFloat(ParquetDictionary.java:41)
    at
org.apache.spark.sql.execution.vectorized.OnHeapColumnVector.getFloat(OnHeapColumnVector.java:423)
    at
org.apache.spark.sql.catalyst.expressions.GeneratedClass$GeneratedIteratorForCodegenStage1.processNext(Unknown
Source)     at
org.apache.spark.sql.execution.BufferedRowIterator.hasNext(BufferedRowIterator.java:43)
    at
org.apache.spark.sql.execution.WholeStageCodegenExec$$anon$2.hasNext(WholeStageCodegenExec.scala:636)
    at
org.apache.spark.sql.execution.columnar.CachedRDDBuilder$$anon$1.hasNext(InMemoryRelation.scala:125)
    at
org.apache.spark.storage.memory.MemoryStore.putIterator(MemoryStore.scala:221)
    at
org.apache.spark.storage.memory.MemoryStore.putIteratorAsValues(MemoryStore.scala:299)
    at
org.apache.spark.storage.BlockManager.$anonfun$doPutIterator$1(BlockManager.scala:1165)
    at
org.apache.spark.storage.BlockManager.doPut(BlockManager.scala:1091)
    at
org.apache.spark.storage.BlockManager.doPutIterator(BlockManager.scala:1156)
    at
org.apache.spark.storage.BlockManager.getOrElseUpdate(BlockManager.scala:882)
    at org.apache.spark.rdd.RDD.getOrCompute(RDD.scala:357)     at
org.apache.spark.rdd.RDD.iterator(RDD.scala:308)    at
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:346)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:310)     at
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:346)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:310)     at
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:346)
    at org.apache.spark.rdd.RDD.$anonfun$getOrCompute$1(RDD.scala:359)
    at
org.apache.spark.storage.BlockManager.$anonfun$doPutIterator$1(BlockManager.scala:1165)
    at
org.apache.spark.storage.BlockManager.doPut(BlockManager.scala:1091)
    at
org.apache.spark.storage.BlockManager.doPutIterator(BlockManager.scala:1156)
    at
org.apache.spark.storage.BlockManager.getOrElseUpdate(BlockManager.scala:882)
    at org.apache.spark.rdd.RDD.getOrCompute(RDD.scala:357)     at
org.apache.spark.rdd.RDD.iterator(RDD.scala:308)    at
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:346)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:310)     at
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:346)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:310)     at
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:346)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:310)     at
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:346)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:310)     at
org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:99)
    at
org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:55)
    at org.apache.spark.scheduler.Task.run(Task.scala:123)  at
org.apache.spark.executor.Executor$TaskRunner.$anonfun$run$3(Executor.scala:411)
    at org.apache.spark.util.Utils$.tryWithSafeFinally(Utils.scala:1360)
    at
org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:414)
    at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
    at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
    at java.lang.Thread.run(Thread.java:748)

错误片段

4

1 回答 1

1

根据文档：

cache() 方法是使用默认存储级别的简写，即 StorageLevel.MEMORY_ONLY（将反序列化的对象存储在内存中）

cache()是一个惰性操作，如果您查看MEMORY_ONLY部分，您会注意到cache()尝试将 RDD/DataFrame 作为反序列化 Java 对象存储在 JVM 中 [一旦您对缓存的 RDD/DataFrame 调用操作]，因此您在 RDD 中的对象反序列化时遇到问题/数据帧。我建议尝试执行一些转换，例如map()检查序列化/反序列化是否正常工作

如果您df.count()在没有任何转换的情况下调用dfspark 并没有反序列化您的对象

于 2021-06-09T19:57:03.500 回答