VAR in Filter issue

[KNW]

Hi,

So I am trying to use VAR in a filter and came across this issue:

Why this:

Phone = COUNTAX (

    FILTER (

        'table',

        ( LEFT ( 'table'[group_name], 20 ) = "XXXXXXXXXX" )

            && ( LEFT ( 'table'[2ndgroup_name], 20 ) = "XXXXXXXXXX" )

            && 'table'[reas_count] = 0

            && 'table'[type] = "Phone"

            ),

    'table'[number])

is not equal to this:

Phone2 =

VAR _LEFT = COUNTAX(FILTER('table', (LEFT ( 'table'[group_name], 20 ) = "XXXXXXXXXX")

            && (LEFT ( 'table'[2ndgroup_name], 20 ) = "XXXXXXXXXX") && 'table'[reas_count] = 0), 'table'[number])

RETURN

COUNTAX(FILTER('table', _LEFT && [type] = "Phone"), 'table'[number])

Only the first DAX measure counts correctly but i would like to use VAR because that DAX measure uses filters from _LEFT few times and I want it to be calculated only once to improve performance. _LEFT is filtering the table correctly but when the second condition [type]="Phone" is added then rows that contain "Phone" are counted without those conditions specified in _LEFT.
Is it even possible to use VAR in a filter?

[Anonymous]

Can you please share the Power bi file?
Based upon your post and my understanding, the variable _LEFT will be storing a number and what you are trying to perform in the 2nd measure after RETURN is that :
You are trying to filter the table ie. Filter( Table, _LEFT && [type]="Phone")
_LEFT is storing a number. So as per me it does not makes any sense in the filter part. If _LEFT is 0, it would be taking false and if it is 1 it would be taking 1. (This is my understanding)

I am not sure about the answer and I might be wrong.

[KNW · Accepted solution]

well that totally makes sense now. 
if var has to be scalar then there is no way it would work.
thanks for having a look into it, that's what i needed.