Working with multiple date keys in table

Hello,

I am having in table A two date keys formatted as YYYYMMDD (CreateDateKey and CloseDateKey).

There is another entity called Calendar with DateKey formatted also YYYYMMDD and Date in date format.

My goal is to calculate Average time to Finish in days which is Average of (CloseDate - CreateDate ---> time taken to finish).

I was able to perform this creating additional columns CreateDate and CloseDate using left&right functions, substracting dates with datediff(measure) and then performing averagex on datediff measure as below. I did not even use Calendar table for this.

Average time to complete a task = 

VAR TIMETAKENTOFINISH = 

datediff(
max('DWH DIM_CLAIM'[ClaimCreateDate]),
max('DWH DIM_CLAIM'[ClaimClosedDate]),day)

RETURN

AVERAGEX('DWH DIM_CLAIM',TIMETAKENTOFINISH)

 Solution above works, but requires adding two date columns based on date keys.
(which are created like 

date(left('DWH DIM_CLAIM'[ClaimCreateDateKey],4),
right(left('DWH DIM_CLAIM'[ClaimCreateDateKey],6),2),
right('DWH DIM_CLAIM'[ClaimCreateDateKey],2)) 

and it creates date column based on date key)

My question is, is there a way to have the same result without adding additional columns (transforming datekey to date - which is probably unefficient)? I found a USERELATIONSHIP function but did not find a solution using it.

Thanks in advance!

Example data Attached:

Table A Lets say Claim Handling:

Table B as Calendar

Expected output is measure that:
>Relates Table A(Claim handling) DateKeys with TableB. 
>Subtract the days between dates.
>Performs Average of subtracted values. 

In this example, output will be like:
Averagex of (CloseDateKey - CreateDateKey) ----> Averagex of (2, 15, 14, 6 , 5, 4, 1), because (Primary Key 1) is (22/01/22 - 22/01/20 is 2, 22/01/20 - 22/01-05 is 15, etc.)

This example output should be averagex(2,15,14,6,5,4,1) = 6,7