I dont think that statement is entirely true. So, its better to take it with a pinch of salt.
Typically, containers are stateless (not 100%). If a container dies, a different container should be able to take its place and operate. Traditional RDBMS systems are single-server monolithic systems with exceptions being the clustered variants of them. If a single-server goes RDBMS running in a container goes down, first of all, it will be unavailable.
In general, all the modern NoSQL databases are distributed in nature with >=2 copies of the data. So, if one node running in a container goes down, the other nodes in the cluster can takeover the operations of the failed container. There will not be unavailability of the system which is very desirable. Moreover, the failed container can be replaced by a new one as and when possible and restore the orginal scale as before. In similar lines, a distributed RDBMS system can also operate smoothly in a container environment.
Having said the above, there are common challenges with both RDBMS and NoSQL systems face in a docker environment. Statelessness is a challenge. Databases are stateful, especially the ones with persistent storage. There is a storage associated with each node of the database. So, when replacing a container with new one, one need to make sure that the correct storage is associated as per the configuration.