We now only have TrieMaps for binary relations, but we want it for n-ary relations. Random access lookup on the first column in n-ary relations with n > 2 is now in $O(size)$, but not in $O(log_{32}(size)$ as it could be.
We have this for binary relations already and this could be added.
Solving this issue does not have to include specialization of algorithms for join and compose per se. That can be done independently.
We now only have TrieMaps for binary relations, but we want it for n-ary relations. Random access lookup on the first column in n-ary relations with$O(size)$ , but not in $O(log_{32}(size)$ as it could be.
n > 2is now inWe have this for binary relations already and this could be added.
Solving this issue does not have to include specialization of algorithms for join and compose per se. That can be done independently.