Is there a more elegant way of using mapreduce to produce a single list containing multiple strings rather than multiple reduces producing individual strings? As an example