This isn't even an either/or proposition. You can inline SQL in pyspark just fine: df = spark.sql("...")