Cube In Tableau

Question

I have few questions for the experts:

Q1- Can we develop a OLAP cube in Tableau? [I know we can develop reports by connecting to relational database and also to OLAP cubes (e.g. Cognos or SSAS). But I am interested to know if we can really develop a cube in Tableau?]

---

Q2- Is there a difference between creating a dynamic dimension in tableau vs having a standalone dimension table? [somebody suggested me to create a de-normalized table and have tableau create the dimension on the fly. but what about records that are missing in the child/fact table. for instance, customer dimension has 10 records while only 8 were exist in the fact table. wouldn't i be missing other 2 if i connect to child/fact table directly?]

---

Q3- What about performance characteristics of Tableau? [I know tableau executes sql statements behind the scene when it displays data in the reporting tool. if i have millions of records in the de-normalized/child/fact table, will it perform fine?]

Thanks, Moiz

Answer 1

Q1. No. Tableau is a visual analytics front-end, not a tool to build a multi-dimensional OLAP store. While Tableau does have it's own in-memory engine, it does not work the same way a cube does (pre-aggregating by dimension and hierarchy).

Q2. Sorry, this question makes no sense to me.

Q3. In the scenario you mention, Tableau's performance is defined by your database's ability to respond quickly. If your database responds quickly, Tableau will be fast. If not, Tableau will be slow. No magic here. In instances where your db is slow, try Tableau's in-memory engine.