python - AttributeError：模块 'apache_beam' 没有属性 'options'

Question

运行 Apache Beam 管道时出现以下错误。完整的错误代码是：

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-12-870f9c2f41e5> in <module>
     39                  file_path_prefix=os.path.join(OUTPUT_DIR, 'ptp-dataset.csv'))))
     40 
---> 41 preprocess()

<ipython-input-12-870f9c2f41e5> in preprocess()
     22       'requirements_file': 'requirements.txt'
     23     }
---> 24     opts = beam.options.pipeline_options.PipelineOptions(flags=[], **options)
     25     RUNNER = 'DataflowRunner' # 'DirectRunner'
     26 

AttributeError: module 'apache_beam' has no attribute 'options'

产生错误的代码是当我尝试调用PipelineOptions该类时。

 opts = beam.pipeline.PipelineOptions(flags=[], **options)
 RUNNER = 'DataflowRunner' # 'DirectRunner'

Answer 1

要解决此问题，pip install请运行最新版本的 apache-beam：

pip install apache-beam[gcp]

重新启动内核，然后使用options.pipeline_options.PipelineOptions. 在本例中，将其更改为：

opts = beam.options.pipeline_options.PipelineOptions(flags=[], **options)
RUNNER = 'DataflowRunner'