I would be glad if you could advice me the right database-technology for my problem:
I want to store a bigger amount of objects (>5 mio.). These objects belong to different types (approx. 2000 categories); and each type has different attributes (approx. 50 per type; 50% of these attributes (not their values) are similar across all types) and some of these attributes can be nested (e.g. in case of ranges: weight has a minimum and a maximum-attributes).
This DB will be extended over the time by more objects. And most likely I need to add more types and adjust attributes (e.g. add/ delete attributes per types).
The queries I want to conduct are: "search for objects by a given string", "filter for objects by given attributes" or "show all objects of a certain type". I'm not (or at least not likely) going to conduct any queries like "give me all objects across all types which have a certain attribute".
I am thinking which DB-technology fits best to my case: a graph-oriented (e.g. Neo4J) or a relational (e.g. MySQL)?
Does anybody have experiences from a similar case? Would be glad to get any hint - from literature etc. it is hard to judge for me and I don't want to start implementing both.
Many thanks in advance!