可以使用以下代码示例在写入后保持分区列作为行值：

在您的脚本中添加以下行：

job = Job(glueContext) job.init('myjob')

Set up logging

logger = glueContext.get_logger()

Set up glue argument parser and parse arguments

glue_parser = argparse.ArgumentParser(description='Arguments for my glue job') glue_parser.add_argument('--s3_destination_path', type=str, required=True, help='S3 destination path where the data should be written') args = glue_parser.parse_args()

Set destination path

s3_destination_path = args.s3_destination_path

Create dynamic frame

input_dyf = glueContext.create_dynamic_frame.from_catalog(database="mydb", table_name="mytable")

Convert dynamic frame to data frame

input_df = input_dyf.toDF()

Write data frame as partitioned parquet files in S3

input_df.write
.partitionBy('partition_column')
.parquet(s3_destination_path)

Read partitioned parquet files from S3 as data frame

output_df = spark.read
.parquet(s3_destination_path)

Drop partition column from data frame

output_df = output_df.drop('partition_column')

Add partition column with corresponding value as row value

output_df = output_df.withColumn('partition_column', lit('value'))

Write final data frame as non-partitioned parquet file in S3

output_df.write
.parquet(s3_destination_path)