python - WriteToText 在 DirectRunner 中工作，但在 DataflowRunner 中因 TypeError 而失败

Question

我可以使用 DirectRunner 运行这段代码，它运行良好。使用 DataflowRunner 它会崩溃:

TypeError: process() takes exactly 4 arguments (3 given) [while running 'write_text/Write/WriteImpl/WriteBundles']`

我的 apache-beam-sdk 是按照说明中的指示从 master 克隆和构建的。它构建为 apache-beam-sdk==0.6.0.dev0。然而，我对版本持怀疑态度，因为(我认为)我最近看到代码更改而版本没有更改(NewDoFn 消失但版本没有更改)。

我不确定这是否是问题的根源，但安装的 sdk 和数据流容器似乎不匹配。我收到另一个不匹配类型的错误，其中 DirectRunner 将 element 直接传递给我的 DoFn.process() 而 DataflowRunner 通过上下文。

我试图将其隔离为尽可能简单的代码:

import uuid
import apache_beam.utils.pipeline_options
import apache_beam as beam

runner = 'DataflowRunner'
# runner = 'DirectRunner'

options = beam.utils.pipeline_options.PipelineOptions()
gcloud_options = options.view_as(beam.utils.pipeline_options.GoogleCloudOptions)
gcloud_options.job_name = 'a' + str(uuid.uuid4())
gcloud_options.project = 'your-project'
gcloud_options.staging_location = 'gs://your-bucket/beam/staging'
gcloud_options.temp_location = 'gs://your-bucket/beam/temp'
options.view_as(beam.utils.pipeline_options.StandardOptions).runner = runner

p = beam.Pipeline(options=options)
(p
 | 'some_strings' >> beam.Create(tuple('asdfqwert'))
 | 'write_text' >> beam.io.WriteToText('strings', file_name_suffix='.txt')
 )
p.run().wait_until_finish()

完整输出:

No handlers could be found for logger "oauth2client.contrib.multistore_file"
/Users/john/miniconda3/envs/py2/lib/python2.7/site-packages/apache_beam/coders/typecoders.py:136: UserWarning: Using fallback coder for typehint: Any.
  warnings.warn('Using fallback coder for typehint: %r.' % typehint)
DEPRECATION: pip install --download has been deprecated and will be removed in the future. Pip now has a download command that should be used instead.
Collecting google-cloud-dataflow==0.5.1
  Using cached google-cloud-dataflow-0.5.1.tar.gz
  Saved /var/folders/v3/61xx4nnn6p36n5m9fp4qdwtr0000gn/T/tmpuCWoeh/google-cloud-dataflow-0.5.1.tar.gz
Successfully downloaded google-cloud-dataflow
Traceback (most recent call last):
  File "reproduce_bug.py", line 28, in <module>
    p.run().wait_until_finish()
  File "/Users/john/miniconda3/envs/py2/lib/python2.7/site-packages/apache_beam/runners/dataflow_runner.py", line 706, in wait_until_finish
    (self.state, getattr(self._runner, 'last_error_msg', None)), self)
apache_beam.runners.dataflow_runner.DataflowRuntimeException: Dataflow pipeline failed. State: FAILED, Error:
(70278eb56b40fd94): Traceback (most recent call last):
  File "/usr/local/lib/python2.7/dist-packages/dataflow_worker/batchworker.py", line 514, in do_work
    work_executor.execute()
  File "dataflow_worker/executor.py", line 899, in dataflow_worker.executor.MapTaskExecutor.execute (dataflow_worker/executor.c:26452)
    op.start()
  File "dataflow_worker/executor.py", line 191, in dataflow_worker.executor.ReadOperation.start (dataflow_worker/executor.c:7575)
    def start(self):
  File "dataflow_worker/executor.py", line 196, in dataflow_worker.executor.ReadOperation.start (dataflow_worker/executor.c:7480)
    with self.spec.source.reader() as reader:
  File "dataflow_worker/executor.py", line 206, in dataflow_worker.executor.ReadOperation.start (dataflow_worker/executor.c:7425)
    self.output(windowed_value)
  File "dataflow_worker/executor.py", line 136, in dataflow_worker.executor.Operation.output (dataflow_worker/executor.c:5749)
    cython.cast(Receiver, self.receivers[output_index]).receive(windowed_value)
  File "dataflow_worker/executor.py", line 83, in dataflow_worker.executor.ConsumerSet.receive (dataflow_worker/executor.c:3884)
    cython.cast(Operation, consumer).process(windowed_value)
  File "dataflow_worker/executor.py", line 505, in dataflow_worker.executor.DoOperation.process (dataflow_worker/executor.c:15525)
    self.dofn_receiver.receive(o)
  File "apache_beam/runners/common.py", line 163, in apache_beam.runners.common.DoFnRunner.receive (apache_beam/runners/common.c:4862)
    self.process(windowed_value)
  File "apache_beam/runners/common.py", line 270, in apache_beam.runners.common.DoFnRunner.process (apache_beam/runners/common.c:7749)
    self.reraise_augmented(exn)
  File "apache_beam/runners/common.py", line 281, in apache_beam.runners.common.DoFnRunner.reraise_augmented (apache_beam/runners/common.c:8108)
    raise type(exn), args, sys.exc_info()[2]
  File "apache_beam/runners/common.py", line 268, in apache_beam.runners.common.DoFnRunner.process (apache_beam/runners/common.c:7660)
    self.old_dofn_process(element)
  File "apache_beam/runners/common.py", line 173, in apache_beam.runners.common.DoFnRunner.old_dofn_process (apache_beam/runners/common.c:5182)
    self._process_outputs(element, self.dofn_process(self.context))
  File "apache_beam/runners/common.py", line 152, in apache_beam.runners.common.DoFnRunner.__init__.lambda3 (apache_beam/runners/common.c:3640)
    self.dofn_process = lambda context: fn.process(context, *args)
TypeError: process() takes exactly 4 arguments (3 given) [while running 'write_text/Write/WriteImpl/WriteBundles']

Answer 1

您的环境似乎安装了 0.5.1 版(请参阅堆栈跟踪的顶部)，但您正在使用 python 存储库的 HEAD 进行构建。

您可以创建一个新的 virtualenv 环境，其中包含正确版本的 SDK。

• 如果您想针对 Python HEAD 运行，则需要在运行管道时设置 sdk_location 标志。
• 如果您想针对已发布的版本运行，请使用 pip install google-cloud-dataflow 安装 SDK，然后正常运行您的管道。 (最好在使用 virtualenv 的虚拟环境中)

请注意，最好使用已发布的版本。