DAX for DISTINCT Count ID on Many to many relationship table

[Anonymous]

Hi All,

I have two tables with Many to Many relation, I am trying to create a column to find the distinct count of "Id" and group it by start and end time. I don't want it as a measure because I am going to use the value from the new column to compare it with another column.

TABLE A

ID ||	Start_Time ||	End_Time ||	15_Min_Time
1 ||	2020-01-02 10:01:00 ||	2020-01-02 10:08:00  ||	2020-01-02 10:00:00
2 ||	2020-01-02 10:01:00 ||	2020-01-02 10:08:00 ||	2020-01-02 10:00:00
3 ||	2020-01-02 10:06:00 ||	2020-01-02 10:12:00 ||	2020-01-02 10:00:00
4 ||	2020-01-02 10:18:00 ||	2020-01-02 10:30:00 ||	2020-01-02 10:15:00

TABLE B

Time  ||	Info_Value ||	15_Min_Time
2020-01-02 10:00:00 ||	2  ||	2020-01-02 10:00:00
2020-01-02 10:01:00 ||	1  ||	2020-01-02 10:00:00
2020-01-02 10:02:00 ||	3  ||	2020-01-02 10:00:00
2020-01-02 10:03:00 ||	4  ||	2020-01-02 10:00:00
2020-01-02 10:04:00 ||	5  ||	2020-01-02 10:00:00
2020-01-02 10:05:00 ||	1  ||	2020-01-02 10:00:00

TABLE A and TABLE B are having M:M relation with 15_Min_Time

 Expected Output:

I want a column with distinct count of id but grouped by Start_Time

ID ||	Start_Time ||	End_Time ||	15_Min_Time ||	New_Column
1  ||	2020-01-02 10:01:00 ||	2020-01-02 10:08:00  ||	2020-01-02 10:00:00  ||	2
2 ||	2020-01-02 10:01:00 ||	2020-01-02 10:08:00 ||	2020-01-02 10:00:00  ||	2
3 ||	2020-01-02 10:06:00 ||	2020-01-02 10:12:00 ||	2020-01-02 10:00:00  ||	1

I tried the following formula

New_Column = CALCULATE(DISTINCTCOUNT(Table A[id]), GroupBy(Table A[Start_Time], Table A[End_Time]))

Error --- Circular Dependency Error

New_Column = SUMMARIZE(Table A,Table A[start_time],Table A[End_time],"Distinct ID",DISTINCTCOUNT(Table A[id]))

Error ---- The expression refers to multiple columns.

[Greg_Deckler]

@Anonymous First, I would highly recommend trying to get rid of any direct many-to-many relationships. Use a bridge table to connect those two tables and your life will be easier.

As for your errors. Your second one is because SUMMARIZE returns a table, not a single scalar value.

The first error, circular dependencies are difficult to resolve without more information. 

Not really enough information to go on, please first check if your issue is a common issue listed here: https://community.powerbi.com/t5/Community-Blog/Before-You-Post-Read-This/ba-p/1116882

Also, please see this post regarding How to Get Your Question Answered Quickly: https://community.powerbi.com/t5/Community-Blog/How-to-Get-Your-Question-Answered-Quickly/ba-p/38490

The most important parts are:
1. Sample data as text, use the table tool in the editing bar
2. Expected output from sample data
3. Explanation in words of how to get from 1. to 2.

[Anonymous]

@Greg_Deckler  Thanks for sharing the information, I will work on resolving the M to M cardinality, I have updated my question with more information, hope this will help in resolving the issue.

[Anonymous · Accepted solution]

I resolved this issue by going through the article https://www.sqlbi.com/articles/understanding-circular-dependencies/ I used the ALLExcept function to overcome the issue.

New_Col =  CALCULATE(DISTINCTCOUNT(Table A[id]),GROUPBY(Table A,Table A[start_time], Table B[End_time]),ALLEXCEPT(Table A,Table A[Calculated_Col]))

NOTE:
Calculated_Col is a column that I created to get the average of Table B(Info_Value), power bi considers all the columns in the table while creating "New_Col", so I was getting a Circular Dependency error, ALLExcept resolves that issue.