I am currently using a Python notebook with a defined schema to import fairly unstructured documents in MongoDB. Some of these documents have spaces in their field names. I define the schema for the MongoDB PySpark connector like the following:

StructField("My Field Name", StringType())

Unfortunately, this gives me the error "Found invalid character(s) among " ,;{}()\n\t=" in the column names of your schema." I would be happy to rename the column, but I have to be able to import it from MongoDB first. Is there a way to do this with the schema? Or am I forced to write a UDF to convert a JSON string with the bad field name into normalized columns?