Feature Request: Explain Predictions: Select the top X attributes by importance for the imp port

It seems there's a missing feature with Explain Predictions: the "imp" output port always returns the importance of all attributes for all examples. I would like to get only the top X for both positive and negative importance for visualization. The "exa" port uses the "maximal explaining attributes" parameter for this, but it does not return the information in a format adequate for the visualization I'm trying to build.

I experimented with a few convoluted solutions to retrieve this list of top X attributes by group but I can't get to a reasonably simple solution.

I was trying to do something similar to this SQL query but did not find a simple solution to implement it:

-- top 5 positive importances
SELECT * FROM table
WHERE row_number() OVER (PARTITION BY GroupingColumn1, GroupingColumn2) ORDER BY (Importance DESC) <= 5
UNION
-- top 5 negative importances
SELECT * FROM table
WHERE row_number() OVER (PARTITION BY GroupingColumn1, GroupingColumn2) ORDER BY (Importance ASC) <= 5

What would be the simplest way to do this in RapidMiner? 

I would also like to create a feature request to add a boolean option "Apply maximal explaining attributes to the imp output port" to the Explain Predictions operator to avoid the need to implement this kind of filtering in the future.