Thanks For the Update Louis... 
As we are planning to Sync All our notebook from Scala to Pyspark , we are in process of converting the code. I think Adding the additional dependency of ABRiS or Adobe’s spark-avro with Schema Registry support will take more regression test for us. But anyways as mentioned from_avro functionality has special overloads with Databricks which is not supported outside it was the bottom line i want to confirm