python - 用户定义的函数破坏了 pyspark 数据帧

Question

我的spark版本是1.3，我用的是pyspark

我有一个名为 df 的大型数据框。

from pyspark import SQLContext
sqlContext = SQLContext(sc)
df = sqlContext.parquetFile("events.parquet")

然后我选择数据框的几列并尝试计算行数。这很好用。

df3 = df.select("start", "end", "mrt")
print(type(df3))
print(df3.count())

然后我应用一个用户定义的函数将其中一列从字符串转换为数字，这也可以正常工作

from pyspark.sql.functions import UserDefinedFunction
from pyspark.sql.types import LongType
CtI = UserDefinedFunction(lambda i: int(i), LongType())
df4 = df2.withColumn("mrt-2", CtI(df2.mrt))

但是，如果我尝试计算行数，我会得到一个异常，即使类型显示它是一个数据框，就像 df3 一样。

print(type(df4))
print(df4.count())

我的错误:

---------------------------------------------------------------------------
Py4JJavaError                             Traceback (most recent call last)
<ipython-input-10-53941e183807> in <module>()
      8 df4 = df2.withColumn("mrt-2", CtI(df2.mrt))
      9 print(type(df4))
---> 10 print(df4.count())
     11 df3 = df4.select("start", "end", "mrt-2").withColumnRenamed("mrt-2", "mrt")

/data/cloudera/parcels/CDH-5.4.7-1.cdh5.4.7.p0.3/lib/spark/python/pyspark/sql/dataframe.py in count(self)
    299         2L
    300         """
--> 301         return self._jdf.count()
    302 
    303     def collect(self):

/data/cloudera/parcels/CDH-5.4.7-1.cdh5.4.7.p0.3/lib/spark/python/lib/py4j-0.8.2.1-src.zip/py4j/java_gateway.py in __call__(self, *args)
    536         answer = self.gateway_client.send_command(command)
    537         return_value = get_return_value(answer, self.gateway_client,
--> 538                 self.target_id, self.name)
    539 
    540         for temp_arg in temp_args:

/data/cloudera/parcels/CDH-5.4.7-1.cdh5.4.7.p0.3/lib/spark/python/lib/py4j-0.8.2.1-src.zip/py4j/protocol.py in get_return_value(answer, gateway_client, target_id, name)
    298                 raise Py4JJavaError(
    299                     'An error occurred while calling {0}{1}{2}.\n'.
--> 300                     format(target_id, '.', name), value)
    301             else:
    302                 raise Py4JError(

Py4JJavaError: An error occurred while calling o152.count.
: org.apache.spark.SparkException: Job aborted due to stage failure: Task 1379 in stage 12.0 failed 4 times, most recent failure: Lost task 1379.3 in stage 12.0 (TID 27021, va1ccogbds01.lab.ctllabs.io): org.apache.spark.api.python.PythonException: Traceback (most recent call last):
  File "/data/0/cloudera/parcels/CDH-5.4.7-1.cdh5.4.7.p0.3/jars/spark-assembly-1.3.0-cdh5.4.7-hadoop2.6.0-cdh5.4.7.jar/pyspark/worker.py", line 101, in main
    process()
  File "/data/0/cloudera/parcels/CDH-5.4.7-1.cdh5.4.7.p0.3/jars/spark-assembly-1.3.0-cdh5.4.7-hadoop2.6.0-cdh5.4.7.jar/pyspark/worker.py", line 96, in process
    serializer.dump_stream(func(split_index, iterator), outfile)
  File "/data/0/cloudera/parcels/CDH-5.4.7-1.cdh5.4.7.p0.3/jars/spark-assembly-1.3.0-cdh5.4.7-hadoop2.6.0-cdh5.4.7.jar/pyspark/serializers.py", line 236, in dump_stream
    vs = list(itertools.islice(iterator, batch))
  File "/data/cloudera/parcels/CDH-5.4.7-1.cdh5.4.7.p0.3/lib/spark/python/pyspark/sql/functions.py", line 119, in <lambda>
  File "<ipython-input-10-53941e183807>", line 7, in <lambda>
TypeError: int() argument must be a string or a number, not 'NoneType'

at org.apache.spark.api.python.PythonRDD$$anon$1.read(PythonRDD.scala:135)
at org.apache.spark.api.python.PythonRDD$$anon$1.next(PythonRDD.scala:98)
at org.apache.spark.api.python.PythonRDD$$anon$1.next(PythonRDD.scala:94)
at org.apache.spark.InterruptibleIterator.next(InterruptibleIterator.scala:43)
at scala.collection.Iterator$$anon$13.hasNext(Iterator.scala:371)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
at org.apache.spark.rdd.RDD$$anonfun$zip$1$$anon$1.hasNext(RDD.scala:743)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
at org.apache.spark.sql.execution.Aggregate$$anonfun$execute$1$$anonfun$6.apply(Aggregate.scala:127)
at org.apache.spark.sql.execution.Aggregate$$anonfun$execute$1$$anonfun$6.apply(Aggregate.scala:124)
at org.apache.spark.rdd.RDD$$anonfun$14.apply(RDD.scala:634)
at org.apache.spark.rdd.RDD$$anonfun$14.apply(RDD.scala:634)
at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:35)
at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:277)
at org.apache.spark.rdd.RDD.iterator(RDD.scala:244)
at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:35)
at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:277)
at org.apache.spark.rdd.RDD.iterator(RDD.scala:244)
at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:68)
at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:41)
at org.apache.spark.scheduler.Task.run(Task.scala:64)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:203)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)

Driver stacktrace:
at org.apache.spark.scheduler.DAGScheduler.org$apache$spark$scheduler$DAGScheduler$$failJobAndIndependentStages(DAGScheduler.scala:1210)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1199)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1198)
at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59)
at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47)
at org.apache.spark.scheduler.DAGScheduler.abortStage(DAGScheduler.scala:1198)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:693)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:693)
at scala.Option.foreach(Option.scala:236)
at org.apache.spark.scheduler.DAGScheduler.handleTaskSetFailed(DAGScheduler.scala:693)
at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1400)
at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1361)
at org.apache.spark.util.EventLoop$$anon$1.run(EventLoop.scala:48)
---------------------------------------------------------------------------

我是否正确使用了用户定义的函数？知道为什么数据框函数对数据框不起作用吗？

Answer 1

从堆栈跟踪来看，您的列似乎包含一个 None 值，它破坏了 int 转换；您可以尝试将 lambda 函数更改为 lambda i: int(i) if i else None，以处理这种情况。

请注意，仅仅因为 df2.withColumn("mrt-2", CtI(df2.mrt)) 没有抛出错误并不意味着您的代码没问题:Spark 具有惰性评估，因此它不会真正尝试运行您的代码，直到您调用 count、collect 或类似的东西。