Memory issues with DAX filter

[anagnostes]

Hi,

I have a model where I have a Customers table and an Orders Table. I want to add a new mesure in the Customers table where i want to compute the average buying cyle. So i add this DAX mesure:

    BUYING_CYCLE = 
             AVERAGEX(ORDERS; 
                     DATEDIFF( CALCULATE ( 
                         MAX ( 'ORDERS'[ORDER_DATE] ) ;
                             FILTER(ORDERS;  EARLIER('ORDERS'[ORDER_DATE]) >= 'ORDERS'[ORDER_DATE] )
                     ); 'ORDERS'[ORDER_DATE]
                     ; DAY )
             )  

This works fine but it's not correct. I should add a second filter like this:

FILTER(ORDERS; EARLIER('ORDERS'[ORDER_DATE]) >= 'ORDERS'[ORDER_DATE] ) && EARLIER('ORDERS'[ID]) <> 'ORDERS'[ID] )

But if I add this second expression the query is really slow and the memory gets to 98 plus %.  There aren't a lot of records in the Orders table, like 25000. 

Can anybody tell me how can i understand what is happening here?

Thank you.

[Baskar]

My understanding from your requierment .

U have customer table, Order Table.

U want to know the Customer average ? am right ?

if yes ? do u have relarionship between this two tables right ?

then u can use direct average function . i know i misunderstood.

Can u please share your sample data, and your expected output i will help u.

I think u want to know the Customer Running Average.

[Bitwize_PowerBI]

I think you can do the filter without the earlier() function.

try: FILTER(ORDERS; VALUES('ORDERS'[ORDER_DATE]) >= 'ORDERS'[ORDER_DATE] ) && VALUES('ORDERS'[ID]) <> 'ORDERS'[ID] )

as you cycle through all of your orders, you could create a calculated column in your orders table to get the 'previous order date'.  You would have to add a condition for 'the same customer':

CALCULATE ( 
                        MAX ( 'ORDERS'[ORDER_DATE] ) ;
                             FILTER(ORDERS; EARLIER('ORDERS'[ORDER_DATE]) >= 'ORDERS'[ORDER_DATE] ) && EARLIER('ORDERS'[ID]) <> 'ORDERS'[ID] && EARLIER('CUSTOMER'[ID]) = 'CUSTOMER'[ID])
                     )

Regards,

Dries