columnar vs cubes

[David Spades]

Hi,

First, perhaps it's better to explain the problem background
I'm tracking 100K+ items sold on various store chains globally. each has item name, store chain name, manufacturer, sales country and date dimension. There are 4 facts that I record --> sales volume, manufactured volume, defect volume and returned volume. Sample of report I want to generate:
* top n items with highest average daily sales volume increase during the whole 2014

I'm trying to do aggregate operations on several millions facts vs decades of time. I just recently stumbled upon columnar DBMS. My question :

1. Are OLAP cubes and columnar DBs partners or competitors in DW?
2. If they're competitors, which is better? there will be no optional columns (sparseness) in my transactional data.

Thanks