It would be useful to have the option, once you've pulled data into DataManager, to nominate a row as the header row. This would mean that you could then manipulate the headers dynamically (with formulae etc).
This idea is relatively popular but it's a little odd. Are you trying to solve the problem where you've imported the data incorrectly and the headers are actually on row 3?
If so this situation is far better resolved by correcting your data at source or by some kind of future "re-headering" data transform/operation.
But if you really do want to replace the displayed field names with a given row's values, only in a specific Table view, please justify this / explain the scenario. Are you trying to recreate the Excel freeze panes functionality?
I know it sounds odd but basically I have run into many problems where I need to manipulate the headings so that DM blocks downstream will work. For example, if you have a pivot block it will pivot on a named field, but if the column heading contains a descriptor and the date, basically the field name changes every time you import an updated file, which then breaks the pivot block.
You could import the file specifying there is no header row, then filter the header row off in one direction where you could use search/replace or formula to clean the headings up and then merge it back together.
That's quite a simplistic example but I hope it clarifies what I was getting at?
So, you are using a row of data for "correct" field names (thus converting all values to text), and are using auto-generated (Field 3...) names to enable consistent de-pivot behaviour?
Please vote, and add further comments to that discussion, describing the different scenarios of de-pivot in terms of the fields you want to select and deselect and how they change over time.