What is the purpose of a fact table in a data warehouse?

[datadonuts]

In the Microsoft Learn section for "Implement a data warehouse with Microsoft Fabric  Query a data warehouse in Microsoft Fabric"

I got this question here Knowledge check - Training | Microsoft Learn

What didn't I understand?

[datadonuts · Accepted solution]

Thanks for the reply. 

If the definition of "raw data" is "meta data, then I agree.

Otherwise, raw data from my understanding are the "original" data from the data source e.g. RDB without any transformations or aggregations, which typically contains the records of a certain business like sales or orders in their lowest granularity. That also means, these tables typically do not contain calculations e.g. price * quantity = net sales, in a column of the fact-table.
Saying this, a fact-table stores raw data about a business and does not contain calculation. Still that's my understanding, but maybe definitions have changed since Christopher Adamson 😉

[datadonuts · Accepted solution]

Thanks for the reply. 

If the definition of "raw data" is "meta data, then I agree.

Otherwise, raw data from my understanding are the "original" data from the data source e.g. RDB without any transformations or aggregations, which typically contains the records of a certain business like sales or orders in their lowest granularity. That also means, these tables typically do not contain calculations e.g. price * quantity = net sales, in a column of the fact-table.
Saying this, a fact-table stores raw data about a business and does not contain calculation. Still that's my understanding, but maybe definitions have changed since Christopher Adamson 😉

[frithjof_v · Accepted solution]

I agree @datadonuts and I answered the same as you on this question.

I have never heard this definition of the purpose of a Fact table before ("To store the results of calculations") and I believe this is not the primary purpose of a fact table.

I would say it's more typical that a Fact table stores events, measurements and/or metrics, and contains foreign keys to the dimension tables which contain descriptive information about the entities involved in each fact instance (=each row in the fact table).

Calculations don't necessarily need to be stored in the Fact table, instead they could be calculated on the fly by the analytical tool (e.g. by using measures in Power BI).
Sometimes, however, we would want to store calculations in the Fact table to ease the processing burden from the analytical tool. But I don't think that storing calculations can be regarded as the purpose of a Fact table.

[Anonymous]

Hi @datadonuts 
Thanks for using Fabric Community.

The purpose of a fact table in a data warehouse is To store the results of calculations.

Fact tables are the primary tables in a data warehouse and they store the measurable, quantitative data that businesses analyze. This data is often the result of calculations. For example, a fact table in a retail sales data warehouse might store data such as the number of units sold and total sales.

Fact tables do not typically store metadata about the data warehouse itself. That information is usually stored in system tables or a data dictionary.

I hope this helps! Let me know if you have any other questions

[Anonymous]

Hi @datadonuts 
We haven’t heard from you on the last response and was just checking back to see if your query got resolved. Otherwise, will respond back with the more details and we will try to help.
Thanks