Skip to main content
Woohoo! Qlik Community has won “Best in Class Community” in the 2024 Khoros Kudos awards!
Announcements
Nov. 20th, Qlik Insider - Lakehouses: Driving the Future of Data & AI - PICK A SESSION
cancel
Showing results for 
Search instead for 
Did you mean: 
Anonymous
Not applicable

Data Modeling in Qlikview - Star Schema vs Snowflake

I have a confusion in choosing the Data Model Schema for my project. Can you guys please guide me choosing the right Schema?

In General , when do we Choose Star Schema over Snowflake and vice versa??

4 Replies
petter
Partner - Champion III
Partner - Champion III

Snowflake is just extending a Star Schema. When it comes to Qlik it seldom makes any difference speedwise unless you have a lot of rows in your dimension tables.

The associative engine in Qlik works equally well for both types.

Anonymous
Not applicable
Author

Thanks for your Answer Petter.

Generally when we are working with any Traditional BI tool People prefer to work with STAR Schema over Snowflake.Reason they give -- Less Joins -->Better  Performance.

My Situation:

Case 1:

6 Dimension Tables (Denormalized) and 1 Fact table ---> My Architect recommended to Normalize the dimensional tables before loading into QV , So that we can have snowflake schema in the data model.

Case 2:

8 Dimensions & 2 facts:

Similar Recommendations.

I don't understand  why he is asking me to Normalize the dimensions .I did ask him about this but didn't get any answer.

petter
Partner - Champion III
Partner - Champion III

Qlik will work functionally similar with both normalized and denormalized dimension tables. Due to compression algorithms in Qlik when it comes to in-memory tables there is little if anything to save of memory by normalizing. So I wouldn't worry too much either way unless normalizing leads to a lot of small tables being created.

In Qlik the "Less Joins -> Better Perfomance" translates to "Less associations -> Better Performance" although practical tests done by experts have shown that there is little performance loss with snowflake models in most cases.

Anonymous
Not applicable
Author

Yes Petter, I agree with you. With the way QV stores the data there wont be much difference in terms of Memory Saved.

My Problem here is Unnecessary processing to have the normalized tables in QV. I want to avoid it and bring the tables as they are in DB (only With required transformations) .

I don't see any advantage with his recommendations.I am trying to understand if there is advantage with Snowflake over Star.