您可以将rdd转换为dataframe并写入json

from pyspark.sql import SparkSession

    spark = SparkSession.builder \
        .appName('SO')\
        .getOrCreate()

    sc= spark.sparkContext

    df = sc.parallelize(
        [('f1', 1), ('f2', 2)]).toDF(["key", "value"])

    df.write.format('json').save('output_path')

json文件中的输出如下所示

{"key":"f1","value":1}
 {"key":"f2","value":2}