Considering MapR-DB is a distributed database, it either belongs to the "CP" category or the "AP" category.
Quoting from http://maprdocs.mapr.com/home/MapROverview/c_maprdb_new.html
When an update is applied to a tablet in the master container (which is at the head of a replication chain), the update is applied serially to the replicas of that container in the chain. The update is complete only when all replicas in the chain are updated.
This implies that in the case of a network partition the replication chain is broken (or delayed indefinitely) precluding operations from completing successfully. This means they chose Consistency over Availability, making MapR-DB a CP system, rather than AP.