When dealing with certain datasets (e.g, a collection of profiles per MPI rank), it can be useful to replace the "profile" index level with data from the metadata table (e.g., MPI rank). Right now, we can sort-of do this with GroupBy.agg(...), but it would be nicer to have dedicated functionality for this.

A new function for this would behave along these lines:

1. Take the name of a metadata column as input
2. Check if all the values in the metadata column are unique (equivalent to checking if every profile has an unique metadata value)
3. Copy the metadata column into the performance DataFrame
4. Replace the "profile" index level with the copied metadata column
5. Update the row index in the metadata table