Equivalent concept of schema from relational databases?



I’m wondering if there is something in SciDB that is the rough equivalent of schemas in relational databases? For example, I have a set of arrays for a project loaded into SciDB. I would now like to start another different, unrelated project - do I need to work in the same “schema” within SciDB or can I setup another one for my new project?



Not (yet).

Adding “namespaces” / “schemas” / “databases” to SciDB is something we’re looking at, along with a number of other improvements to the logical model: users/roles/privileges, DML to modify arrays, better client/server connectivity, for examples.

But just to set some expectations, we’d much rather spend the time/energy on improving the core functionality. For example, being able to add new physical compute resources to a (running) SciDB instance is something we want to tackle before we get to “namespaces” / “schemas” / “databases”.

Anyone got any thoughts on these priorities?


My personal preferences & reasoning are:

  1. more linear algebra / numerical methods - this is where SciDB should be the clear leader over everyone else

  2. namespaces/schemas/databases - nice to organize things logically rather than install something new

  3. adding physical resources to a running instance - pretty cool but I can just shut things down for awhile. But I can understand the need for this