Filters and structuring for end users

[kazlik]

I have a client who will be customzing content packs and adding/editing charts. They also would like filters which group the same data in different groups, last 30 last 60 days or group A (1,2,3) and group B (1,4,5).

Is it best to have this data repated in a table, even the dates in something like below so they could add and remove these as filters and then just select Group A or B or a more complex option with measures.

Group ID

A      1

A      2

A      3

B      1

B      4

B      5

[Greg_Deckler · Accepted solution]

So, if I understand what you are asking, you have this data in a table already and you are asking whether you should repeat the information in another table and then relate the tables to one another? If that is the goal, then you have to keep in mind that one of the tables needs to have unique values. So, if in your data you have what you show below along with some additional information, you would need to have your "filter" table just have:

Group

A

B

And then you would relate the two group columns.

Now, whether you *need* to do this or not is somewhat questionable and sort of depends on your data and data model. Power BI Slicers and Filters are pretty smart. Slicers already filter to unique values of data for example. In my opinion, this sort of makes some of the classic star schema way of building data models sort of obsolete in many cases for Power BI data models, you are just adding useless complexity to the data model in a lot of cases when you can just have a fact table and some slicers for your "dimensions", but that is purely my opinion.

[kazlik]

Thanks for your input, in the past I would hide a lot of the logic in Measures when using Excel. Now with more interaction with the data, a good thing I am reconsidering how to model things.

I have 10 plus measures and could then have varients for each like this

Volume Group A

Volume Group B

Volume Group C

or I have a seperate table with repeated data and the group association.