Having lots of issues flattening a hierarchical dataset in a Flatter hierarchy.

I am trying to catch instances when there is a successful order import, but a comment on the line item (OE0180.COIN) was truncated.  After the order import, I need to flatten a hierarchical dataset which has OE0520 - OE0500 - OE0180.  In our case, there can only be one comment which will be truncated, so for every OE0520 record, there will be only one OE0500 record, which can only have one OE0180 record, so the hierarchy is always 1:1:1.

The first time I try to open the Flatten transform, I am told that there are duplicate fields (OE0520_BILCITY).  This is incorrect in both the UI as well as the job XML.  Because I cannot see or process any data when the transform is initially opened, my only other option is close the transform and try again.

When I re-open the transform, the "Field Mapping" screen is populated, and I can process the data and it shows up on the right. Good times.  But then when I try to set the "Import" flag to false so I can remove unnecessary columns, my changes aren't reflected in the "Preview" pane, and the field names all changed from their shortnames (BILCITY, BILCOUNTRY) to their full names (OE0520_BILLCITY, OE0520_BILCOUNTRY).

At this point, the dataset becomes un-modifiable.  Any changes I make in the "Field Mappings" pane are not reflected in the "Preview" dataset. I can still extract the fields that I need out of the dataset for my purposes, but it may hinder other users with a slightly different use case.