Question

I’m reading a Wikipedia article about OLAP and OLAP Fact Tables and the article states

All foreign keys between fact and dimension tables should be surrogate
keys, not reused keys from operational data.

But it doesn’t state why.
So, why should

All foreign keys between fact and dimension tables should be surrogate
keys, not reused keys from operational data.

?

Answer 1

From the detailed explanation of what is a Dimension Table from wikipedia:

It is recommended that the key field is a simple integer for the reason that key value is meaningless and is only used to be join fields between the fact and dimension tables.
The usage of surrogate dimension keys brings several advantages among:

• Performance – join processing is much more efficient if a single field surrogate key is used,
• Buffer from operational key management practices – prevents form situations when removed data rows might reappear when their natural keys might be reused or reassigned after a long period of dormancy,
• Mapping to integrate disparate sources,
• Handle unknown or not applicable connections,
• Track changes in dimension attribute values.