MSTR questions 1

[iolaper]

I guess I ve just overlooked these things till now but here they are:
1) We use keys as IDs in attribute definition but sometimes we use dates or text too as IDs. They are obviously not unique. So why is that? I realize though that for example Hiredate (datetime) or Contactdesc (text) while being defined as IDs will be on ONE AND ONLY ONE lookup table. Any other thoughts...

2) Hierarchies are used to define attribute relationships and order while user drills on reports. Is there any other purpose of hierarchies?

3) Lets assume we have a column called Invoicedate. We can define it both as an attribute as well as a metric. In what scenarios do we determine the definition.
I know we define metrics to perform aggregations etc. but any other reason?

Guys I am sorry to bother you with such 101 level questions, but just that it will help me understand microstrategy better.

Thanks a million!!!

 

[entaroadun]

MSTR is a ROLAP platform, so think in terms of SQL.

In ROLAP, every query is a GROUP BY query, so every field must either be in the GROUP BY clause or have an aggregation function.

Attributes are in the GROUP BY clause. Metrics (and facts) have an aggregation function. That is the only difference. Forget DW theory. The difference comes down to SQL.

As for hierarchies:

MSTR maintains attribute relationships in a System Hierarchy object, which it refers to for navigating attribute relationships. You can't alter this. User-defined hierarchies are only good for drilling, etc. In other words, a hierarchy is necessary for MSTR to work, but you can't touch it. The hierarchies you can touch are trivial.