qusetion about context overwrite

[ryan-gao]

my question:

The evaluation of filter context happens and came into data model with its final result ,after that the first parameter of calculate was evaluated.

so even if  “apple, electric”  in test3 has no product, because in this step filter context is not yet came into model, it was immediately overwrite to "tv,electric" in the calculate function,  i think it should return the same result as test1

Am i miss something? wish you can understand my description

thanks

[Phil_Seamark]

HI @ryan-gao

Your pivot table in Test3 is showing a 1 in the cell for TV and Electric, so it's looking fine.

[ryan-gao]

@Phil_Seamark thanks for the reply ,what confused me  is the red box of the pivot table, i think it should also return 1 

[Phil_Seamark]

HI @ryan-gao

Your dataset consists of three rows and you have no rows with both Apple AND Electric.  Filter context is accumulative and applies a logical AND, so only rows that meet every condition will be returned to the COUNT function.

[ryan-gao]

Thanks @Phil_Seamark

as you said,if  the filter context is accumulative, then the left measure test1 's outer filter context is same with the red box on the right,which stimulates test3 's pivot environment.

although  "apple, electric" return no rows, test1 's filter context was overwrite to "tv , electric" by inner filter context. why this can happens on test1 , not test3,  I‘m a bit confused

thanks

[ryan-gao · Accepted solution]

thanks to @Phil_Seamark and @Marco

i think the answer is AutoExists:"the engine performs a first step in which it determines existing
combinations and only later it computes the effective values"  this will happen before the evaluation of filter context