I would like to learn more about how to integrate NoSQL databases to an architecture centered on the relational model (build according to Data Vault 2.0 Standards). Does anyone have an idea of where I could educate myself on the subject. This is currently not in Dan Lindstedt's most recent book!
-
Is there any special requirement you would like to meet or is it more of a general question? – tobi6 Sep 09 '16 at 09:19
1 Answers
I assume with "how to integrate NoSQL databases", you are referring how you would use a NoSQL data source as an input into a relational DV2.0 model. Simply put, you just need to ingest the data into your relational database and transform the data into the required DV structures according to your business entities (hubs), their attributes (satellites) and their relationships (links and satellites). This also assumes that any unstructured data you have in the NoSQL database is either out of scope or can be stored in the relational engine's BLOB, XML or JSON store.
If you are referring to a DV2.0 modelled and stored in a NoSQL database then you are right, this is not covered in Dan's new book. Although the DV2.0 standard should be easy to apply to something like Hive, Cloudera and Cassandra. The hashes make it possible to load every entity completely separately from the other one, complex values (unstructured or semi-structured data) can just be stored as attributes on the satellite. How you compare those value for the next load depends on the data. Perhaps, a process of storing a SHA1 hash alongside the unstructured data would be useful when doing a comparison.
Good luck and let me know how you get on.

- 21
- 1
-
If we import data from NoSQL to the DV model and store them into satellite we are not exploiting the capacity and the advantages of NoSQL databases. For example, if I want to use Neo4j's graph algorithms, there is no point of storing the relationship of the graph in links... If we are moving from NoSQL to SQL based solutions, we are not integrating but converting. This does not seems to be right. – user2058291 Sep 09 '16 at 12:31
-
Not exactly, "how to integrate NoSQL databases", I want to use both types of database in a system in harmony! Not import data from one database to another. I want to store what is graph oriented in a DB such as Neo4j and what is relational in DV and what is document-oriented in MongoDB etc... I want to use them jointly! – user2058291 Sep 09 '16 at 12:49