The definition of an operational data store is a data store, which is used for operational purposes. This data store could be getting updated online real time basis or could be done on offline basis (say once a day refresh). Operational data store is one among many applications of a Data Warehouse. We have mentioned in another question, that data warehouse can have multiple applications. A Data Warehouse can also act like an operational data store, if the dimensional model etc. is structured accordingly. It is also possible to have an operational data store separate from a data warehouse. For example, if data warehouse is also acting like an operational data store, it will need include some additional data (like detailed customer address), which otherwise, it might not have included.
Another question one may ask is that will the approach of creating an ODS and Data Warehouse be different?
The answer to this will be- Yes. If a data warehouse is going to be used as an operational data store as well, or you are creating an ODS separate from a data warehouse, here will be the differences:
- The non-analytics data will be more extensively included. For example you would like to have the complete address of the customer (instead of pin code, state or city only).
- The data modeling will be more relational. This is because you are running single record, OLTP kind of queries and not the kind of 'mass calculation and aggregation' queries.