要使用Apache Beam Spark Portable Runner将Beam代码在Spark上运行，可以按照以下步骤操作：

1. 导入所需的库和模块：

import apache_beam as beam
from apache_beam.runners.portability import spark_runner

2. 创建一个Beam Pipeline对象：

with beam.Pipeline(runner=spark_runner.SparkRunner()) as p:
    # Beam pipeline code goes here
    ...

3. 在Pipeline中添加所需的转换操作：

with beam.Pipeline(runner=spark_runner.SparkRunner()) as p:
    lines = p | 'ReadLines' >> beam.io.ReadFromText('input.txt')
    words = lines | 'ExtractWords' >> beam.FlatMap(lambda line: line.split(' '))
    word_counts = words | 'CountWords' >> beam.combiners.Count.PerElement()
    word_counts | 'WriteCounts' >> beam.io.WriteToText('output.txt')

4. 运行Pipeline：

with beam.Pipeline(runner=spark_runner.SparkRunner()) as p:
    ...
    p.run()

这样，你就可以使用Apache Beam的Spark Portable Runner在Spark上运行Beam代码了。请确保已经正确配置了Spark环境并安装了相关的依赖库。