Data Modeling - Dimension Date

[DavidPorTodos]

Hi,

I have 3 fact tables with 2 different Date type

Date Fact 1 and 3		Date Fact 2
01.01.2000		2016 Q1
01.01.2000		2016 Q2
01.01.2020		2016 Q3
30.01.2017		2016 Q4
30.01.2000		2017 Q1
31.01.2017		2017 Q2
07.02.2000		2017 Q3
28.02.2020		2017
01.03.2000		2018

The user is asking to being able dynamically to filter the date by year then through a second filter by quarter.

I am looking for the best data model suitable for the case and the creation of Dimensions who could help me in achieving the user goal.

I wonder if i should create a DimDate with all dates day by day, year column and quarter column but the result would be in the case of Fact table 2 a many to many relationship or 2 Dimensions, year and Quarter, after extracting those from all fact tables and creating in each table the corresponding columns in order to relate them.

Thanks for your advice

[DavidPorTodos]

Many thanks for your advice.

[DataInsights · Accepted solution]

@DavidPorTodos,

The approach I take is to add a date column to the fact table. This allows you to create a relationship with DimDate.

To create a date column in the fact table, you can use any date in the period. I prefer to use the first date of the period since all months have day 1, but not all months have day 31. For example, "2016 Q1" would be 2016-01-01 (assuming your quarters are based on a calendar year).