Is a cube database, simply an RDBMS—such as PostgreSQL or MySQL—configured for OLAP via a star or snowflake schema? Or is a cube database different from an RDBMS, and if so how?
Cubes are very a different beast from a traditional database. There are several different kinds of cube storage processes depending on need (OLAP, MOLAP, ROLAP, etc.) which are all done differently depending on how real time the data needs to be.
I actually did a webcast with another Microsoft MVP a couple of weeks ago where we talked a little about the BI process and how it impacts the traditional database, which might be interesting to you. The recordings (which are free) aren't available yet, but should be soon.
-
Can we view the raw data in a cube? Could we see the detail orders of each customer in an ORDERS cube? – mingchau Jun 25 '19 at 6:42
The short answer is no.
Generally an OLAP database will precalculate aggregations while an OLTP system will have to do those calculations when it is asked.
The language for interacting with cubes is appears similar to SQL but MDX is far different beast.
Simplistic answer: A cube allows you to define dimensions (things you'd need to filter on like Dates, , regions,....), hierarchies (eg months within a year, countries for Continents,...) and measures and rules on how these measures relate with the dimensions... Then you can query values based on that using MDX
More simplistic view?
An RDBMS would be an spreadsheet and a cube a Pivotable...
Hope this helps
-
Likening RDBMSs with spreadsheets is a guarantee for 0 upvotes in this site ;) – ypercubeᵀᴹ Nov 9 '13 at 20:39
-
Sure, this metaphor is very approximate, but it makes a cube tangible for someone, who is new to it. Also, I would say, that on the surface a cube is just one more layer of abstraction over an RDBMS. – user77376 Dec 18 '14 at 10:47
