Forum Discussion
Excel DAX query with bridge table using CROSSFILTER
JFGrenier Sure, that's the best decision, you may write a complex DAX that might break tomorrow if not written considering every possible detail, but a data model that's perfect would always give you the required soution with minimal DAX!
Community ChampionJFGrenier Just change the data model a little bit and the DAX that you have to write will be short and simple, working PBI file is attached below my signature:
Hello AntrikshSharma!
And thank you for your reply!
There's a specific reason for the table CountryGroup. Sorry for not mentionning it before. It's because a country can be in multiple country group. For example: Egypt can be in the country group Africa and also in the country group North Africa. What you propose would create a many to many relationship.
Thank's again!
- AntrikshSharma
JFGrenier Still you should focus on changing the model, create 2 rows for Egypt in that case, with a unique key for both?
Let's say that Egypt is in 3 CountryGroup, thing is that I would have to copy the data associated with Egypt 2 more times with a different Id. I'm already at 14 millions rows in the DataTable and for a 32 bits system, I think it's near maximum. But what if I add multiple Id column in my CountryGroup table. I'll check how many country can be in multiple CountryGroup. If there are 15 country, I'd just add 15 columns of Id's and change the relationship accordingly. Do you think the DAX function RELATED could work? Or do you know of any other function that can change relationships? My DAX queries are built at run time via an Excel userform and trapping these exceptions would be easy. You point me in a new direction! Changing the model maybe my only solution!
But before that, I'll give one more shot at this query:
Where (syntax) could I add a filter to the DataTable[Year]=2019 && DataTable[HS2]=10 in the query below?
In the Summarize function, in the Calculate or else???
EVALUATE ADDCOLUMNS ( SUMMARIZE ( DataTable, DataTable[Year] ), "Val", CALCULATE ( SUM ( DataTable[Value] ), CROSSFILTER ( DataTable[IdCountry], Country[IdCountry], BOTH ), CROSSFILTER ( Country[IdCountry], CountryGroup[IdCountry], BOTH ), FILTER ( CountryGroup, CountryGroup[IdCountryGroup] = 1 ) ) )Thank's again!
CNENFRNLI personally avoid leveraging CROSSFILTER as much as possible, especially in a complex data model.
As to your model, in order to propagate filters on CountryGroup(*) to Country(1), expanded table is an excellent choice,
CALCULATETABLE ( VALUES ( Country[IdCountry] ), CountryGroup )Furthermore, filters on Country(1) or Month(1) propagate in a natual way to Data(*) for any calculations; thus, I'd author a measure this way,
Total Values = CALCULATE ( SUM ( Data[Value] ), CALCULATETABLE ( VALUES ( Country[IdCountry] ), CountryGroup ) )