File reference is created when converting from CSV to Parquet

I have a dataflow that converts a csv file to parquet.

Here is my flow:

The issue appears to be in the sink. I want the file to be added in a folder that is in a container so I have this in my dataset:

When I run this, the folder is being created and the parquet file is being added to the folder but then outside of the folder is a random file that is being created with 0 bytes and has the same name as the folder path shown in the dataset. This file has no content type and is essentially blank. When I download it it's actually the folder referenced in my dataset.

This also happens when I select to move my source file to a new folder: