Forum Discussion
join tables vs relation between tables
Xue's formula has a flaw in it at the total level, because it is filtering 'Table' on MAX(Table1[ProductID] ) at the grand total level the max is equal to 2. So the SUM(Value) returns 650 and the SUM(Table1[Allocation]) is not filtered at all as the filter is applied to the other table, so it returns the sum of all Allocations which is 200%. So 650 x 200% = 1300
For logic like this you would be better to use a SUMX pattern to loop over one of the tables row by row. (I'm also filtering on both date and product as per your original requirement)
Allocated Value = SUMX( 'Table 1', Var _date = 'Table 1'[Date] var _product = 'Table 1'[Product] return 'Table 1'[Allocation] * CALCULATE( SUM('Table'[Value]), 'Table'[Date] = _date, 'Table'[Product] = _product) )
Hi andresp_g ,
>this is the expected result by line, however the total units doesnt seem to be equal, do you know how to fix it?
For the sample data, the date is always 1. It can't be shown as a continuous line chart. Suggest you to try the measure and add the columns into Line chart based on your actual scenario. And which column do you mean for total unites?
>since there will be several products and days, how should be the relation between the two tables, should be a combined key made by date & product? or tables are not related?
For my sample, there is no relationship between the tables. You could try it firstly without relationship and check if it is what you want. And if there is any other calculation in your report, maybe you should create it.
>if the inventory table would have other columns as the color column, should be added in the filter parameter?
If you just what the result of the measure, you don't need filter. If you want to view the result of different color, you could add a color slicer to filter them.
(I don't understand your actual report, so probably can't reply you acurately. )
Best Regards,
Xue Ding
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
Xue's formula has a flaw in it at the total level, because it is filtering 'Table' on MAX(Table1[ProductID] ) at the grand total level the max is equal to 2. So the SUM(Value) returns 650 and the SUM(Table1[Allocation]) is not filtered at all as the filter is applied to the other table, so it returns the sum of all Allocations which is 200%. So 650 x 200% = 1300
For logic like this you would be better to use a SUMX pattern to loop over one of the tables row by row. (I'm also filtering on both date and product as per your original requirement)
Allocated Value = SUMX( 'Table 1',
Var _date = 'Table 1'[Date]
var _product = 'Table 1'[Product]
return 'Table 1'[Allocation] * CALCULATE( SUM('Table'[Value]), 'Table'[Date] = _date, 'Table'[Product] = _product) )