This is not an issue but probably a script adjustment.
Would the performance of the script improve if we would read the parquet files with Glue API:
input = glueContext.create_dynamic_frame_from_options("s3", connection_options={"path": path}, format="parquet", transformation_ctx="input").toDF().withColumn("Op", lit("I"))
This is not an issue but probably a script adjustment.
Would the performance of the script improve if we would read the parquet files with Glue API:
input = glueContext.create_dynamic_frame_from_options("s3", connection_options={"path": path}, format="parquet", transformation_ctx="input").toDF().withColumn("Op", lit("I"))
Instead:
input = spark.read.parquet(path).withColumn("Op", lit("I"))
Thank you.