SQL to DAX

[kamgib]

Two tables, Table A has User/Event/EventDate. Table B has User/InteractionDate.

I would use the following SQL Statement to get a one to many table; 

     Select A.User, A.Event, A.EventDate, B.User, B.InteractionDate
      From A, B
      where A.User = B.User
             AND
                (
                  B.InteractionDate >= A.EventDate
                         AND
                  B.InteractionDate <= DateAdd(A.EventDate,30)
                 )

the A.User = B.User is related and can be in a JOIN instead of in the WHERE Clause.

any help to convert this to a DAX or PowerQuery process and get the same results?

[v-rzhou-msft · Accepted solution]

Hi @kamgib ,

Here I create a sample to have a test.

A:

B:

By Power Query:

let
    Source = Table.NestedJoin(A, {"User"}, B, {"User"}, "B", JoinKind.LeftOuter),
    #"Expanded B" = Table.ExpandTableColumn(Source, "B", {"User", "InteractionDate"}, {"B.User", "B.InteractionDate"}),
    #"Filtered Rows" = Table.SelectRows(#"Expanded B", each [B.InteractionDate] >= [EventDate] and [B.InteractionDate]<=Date.AddDays([EventDate],30))
in
    #"Filtered Rows"

 Result is as below.

By Dax:

Dax Table = 
VAR _ADD = ADDCOLUMNS(A,"B.User",RELATED(B[User]),"B.InteractionDate",RELATED(B[InteractionDate]))
VAR _FILTER = FILTER(_ADD,[B.InteractionDate]>=[EventDate]&&[B.InteractionDate]<=[EventDate]+30)
RETURN
_FILTER

Result is as below.

Best Regards,
Rico Zhou

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[kamgib]

thanks for posting, I will give it a try and report back.

[v-rzhou-msft]

Hi @kamgib ,

Could you tell me if your issue has been solved? If it is, kindly Accept the helpful reply as the solution. More people will benefit from it. Or you are still confused about it, please provide me with more details about your data model and your issue or share me with your pbix file without sensitive data.

Best Regards,

Rico Zhou

[PAG]

Hi,

I think you can in PowerQuery merge the two tables

then create a column that adds 30 days to the Eventdate column, call it Eventdate30 for example

Next create a column that subtract the column InteractionDate by Eventdate30 and another that subtract InteractionDate by Eventdate.

Now you can filter the merged table by the last two columns you've created like you want

Hope it helps

PAG