I would use something like:
(let [m [[1 2] [3 4]] max (emax m)] (filter #(= max (apply mget m %)) (index-seq m)))
Would be more efficient to do this with something like ereduce-indexed but we haven't implemented that yet 🙂
yeh I'm using index-seq but that's linear & I was hoping for a logarithmic solution - like, if the matrix were stored as a tree
your fn is shorter than mine though so I'll use it 🙂
i was doing the reduce - though interestingly according to criterium they're almost exactly the same