The ultimate Fabric, Power BI, SQL, and AI community-led learning event. Save €200 with code FABCOMM.
Get registeredCompete to become Power BI Data Viz World Champion! First round ends August 18th. Get started.
Hi everyone,
I have a question concerning the modelling conception.
I have the following model which represent a part of my real model
The tab CATEGORY is created to give more information (with "Details") about the category of the model of car.
CS1 and CS2 belong to the Cat1 which is : aaaaaaaa bbbbbbbb cccccc
Actually, in my PIBX there is any link between these 2 tables and a measure has been created to display the details of the category WHEN THE USER CLICK ON ONE OF CATEGORIES in the report page.
I would like to know if this approach is good , in my opinion I don't think so but I miss some explanation. What do you recommend to do to get the best way to implement a good modelling conception ? If you have some article talks about this subject, It could be interesting too 🙂
Thank you !!!
you would need to put in more data to understand the model more effectively, the model really depends on what you want to do with it, what is your business requirement.
as a side point, what is the value to having category separate, is category unique to the car, or does it apply to different models?
Proud to be a Super User!
Sorry it's my bad , it's the data model.
In fact there is only distinct category in the tab Category and 1 category can have N car.
But I can not set 1...*(1 Category ... N Car) because in the report page, we are able to click on one of differents categories (data of the tab car) and it will display the details of the category choosen.
I don't know if the approach is correct, can you juge please ?
Hi,
According to your description, i advise you not to use many-to-many relationship.
Reference: https://amybayer.wordpress.com/2013/12/29/many-to-many-marketing/
And i advise you to create another dimension table just like this:
Dimension Table = DISTINCT(SELECTCOLUMNS(Car,"Category",Car[Category]))
Then create one-to-many relationships among these three tables:
Best Regards,
Giotto Zhi
Could I know why you don't advise me to not use many to many ? Sorry I didn't have time to read your reference yet.
Thank you for the suggestion and I think It's redundant to have one more information about the category.
In fact, currently, the previous developer has created a pie chart with only data of the tab CAR. That's why the link is * ... * to be able to have the cross filter direction BOTH. Then, when the user clicks on the category of the report , the details of this category will be displayed. It's like this
I just had the idea to do like this. It allows me to help the link 1 ... * ( 1 Category ... * Car ) and it satisfy always the requirement of the user.
What do you think ?
Thank you
Hi,
I am confused about your requirement, do you just want to show details of each category?
If so, please create relationships as this:
And then add [Details] to pie chart visual, it shows:
Best Regards,
Giotto Zhi
In fact, my client want to when he click on one of the category of the pie chart, a sentence (content of "Details" of Category) will be displayed in the same page of report and not in the tooltips of the pie chart.
It's advised to avoid the cardinality *...* because the table category has only unique values. Do you agree ?
Can I ask you why did you create a new table Dimension Table ? what is the aim ?
Hi,
Please try to create a table visual and put [Category] and [Details] into it, when you click on the pie chart, it shows:
Best Regards,
Giotto Zhi
Hi Vanessa,
Thank you for your answer and I'm sorry if I was not clear in my description.
In fact, the model was already like this and I have to improve the modelling to optimize the performance. So I don't know why the previous developer has done like this.
The aim is display the details of the category when I click on one of the category in the report page
As you can see in my photo, a category can include severals differents models