Similar to the Java implementation, we should update the existing name_mapping on update_schema().

Some changes:

• Table.name_mapping no longer defaults to the auto-generated name_mapping based on the current Schema. We should explicitly return None if the property isn't set, so that we can use the function return value to avoid the additional steps in updating the existing name mapping.
• I've introduced some changes compared to the Java version. Specifically:
  • the way it was handling reassigned names had a bug where it wasn't removing all reassigned fields, but only the last one: https://github.com/apache/iceberg/blob/main/core/src/main/java/org/apache/iceberg/mapping/MappingUtil.java#L161C11-L161C19
  • remove MappedField from MappedFields if the mapped_field.names parameter is an empty list