Hi
1: I am reading a parquet file from AWS s3 storage using spark.read.parquet(<s3 path>) 
2: An autoloader job has been configured to load this data into a external delta table.

3: But before loading into this autoloader I need to do some typecasting of data types to match them with my target table.
I have tried using of "schemaHints", tried casting of data after loading data and pass the transformed DF to autoloader as source, "inferSchema" and mergeSchema options to do this. 
But none of them are not working. The casted columns are storing in '_rescue_data' directly.

Is there any way to do this?
NOTE: I need to avoid maintaining multiple locations ie., avoid writing transformed data to a another parquet file and read it from their.