Ingest several types of CSV's with Databricks Auto Loader

I'm trying to load several types of csv files using Autoloader, it currently merge all csv that I drop into a big parquet table, what I want is to create parquet tables for each type of schema/csv_file

Current code does:

#Streaming files/ waiting a file to be dropped
spark.readStream.format("cloudFiles") \
  .option("cloudFiles.format", "csv") \
  .option("delimiter", "~|~") \
  .option("cloudFiles.inferColumnTypes","true") \
  .option("cloudFiles.schemaLocation", pathCheckpoint) \
  .load(sourcePath) \
  .writeStream \
  .format("delta") \
  .option("mergeSchema", "true") \
  .option("checkpointLocation", pathCheckpoint) \
  .start(pathResult)

What I want