1

I followed kimball approach but the problem is I didn't make data marts before star schema. Now my question is can I make data marts in SSAS or is it mandatory to make these marts before schema in kimball approach?

More Information: I am building a warehouse for retail store in my FYP. ![enter image description here

[enter image description here]

Amira Bedhiafi
  • 8,088
  • 6
  • 24
  • 60
zara rana
  • 59
  • 6
  • 1
    A datamart is a logical group of tables, a star schema is 1 or more fact tables and their associated dimension tables. Therefore your question makes no sense. Please can you clarify what the problem is that you are trying to address? – NickW Apr 02 '21 at 16:41
  • In question I added picture of kimball structure. Now my question is is it okay if I didn't make data marts before star. I mean I didn't followed kimball approach completely. – zara rana Apr 02 '21 at 18:11
  • 3
    Your datamarts are on the wrong side of the data warehouse in your diagram, anyway. Datamarts are logical/physical subsets of the data warehouse, not a predecessor to it – NickW Apr 02 '21 at 18:36
  • 2
    I agree with @Nick - what you call "Data Marts" look more like staging databases for the data extraction. Data Marts are subsets of the data warehouse serving specific applications or business units (i.e, Financial Data Mart, Marketing Data Mart, etc). – RADO Apr 02 '21 at 19:23
  • Thanks also is it mandatory to make data marts in kimball approach or I can make it in analysis side. As the difference I studied between kimball and inmon is that in kimball we make data marts in data side whereas in inmon we make data marts in analysis side? – zara rana Apr 02 '21 at 20:41
  • Please can you explain what you mean by "data side" and "analysis side"? Also, what is your definition of a data mart - as this is a term that people use in different ways? – NickW Apr 03 '21 at 10:27
  • What i know is data mart are the subset of data warehouse. And can be used either in analysis or in all data like i post a new picture above which states in kimball approach we use data mart for all data and in inmon we use at analysis. Analysis means in SSAS for queries and for all data means in warehouse. I am beginner so i am confused i didn't make data mart is it make any difference? When my teacher ask me where are your data mart which i showed them on kimball structure what would i say. So i want to know is it mandatory part? – zara rana Apr 04 '21 at 11:20
  • 2
    What the image you've posted actually says is that Kimball uses dimensional modelling for everything whereas Inmon uses dimensional modelling only for data marts - you seem to be confusing the 2 terms "dimensional model" and "data mart". A data mart is a logical subset of a data warehouse, not something you physically build. In my opinion, a data warehouse should be capable of supporting any end user tools but if you are only using SSAS then you could, in theory, build the bare minimum in the DW that SSAS requires and build as much as is possible in SSAS – NickW Apr 04 '21 at 15:37

2 Answers2

0

Data marts should be at the far right of your diagram: specific databases (or cubes!) aimed at a specific use case or audience.

You absolutely don’t need individual data marts before the main data warehouse and the Kimball approach doesn’t suggest you do.

Both Kimball and Inmon can be used for an all in one data warehouse, and then you have the option of making data marts from them, but it is not mandatory.

Rich
  • 2,207
  • 1
  • 23
  • 27
-1

The recommended practice for Kimball's approach is to design and create data marts before implementing the star schema in the data warehouse. However, if you have already implemented the star schema in your data warehouse without creating separate data marts, it is still possible to create data marts using SSAS (SQL Server Analysis Services).

Just keep in mind that data marts are subsets of the data warehouse that are focused on specific business areas or subject areas. They typically contain aggregated and pre-calculated data optimized for analysis and reporting purposes. While it is ideal to design data marts upfront in the Kimball approach, you can still create them after implementing the star schema.

In SSAS, I used to define separate cubes or tabular models to represent the data marts. You have the possibility to model the dimensions and measures specific to each data mart and populate them with data from the existing star schema tables in your data warehouse.

In contrast to the Kimball approach, the Inmon approach to data warehousing emphasizes the creation of a centralized enterprise data warehouse (EDW) that serves as a comprehensive and integrated repository of data from various sources. In the Inmon approach, the focus is on building a normalized data model that captures the entire range of business data across the organization.

Amira Bedhiafi
  • 8,088
  • 6
  • 24
  • 60