Problem with GROUP BY for Duplicates

[imani_tech]

The source table has 3 columns:

[User]
[Delegate]
[Permissions]

The three columns combined should create a compound natural key, but I have found some duplicates using T-SQL;

SELECT
  [User],
  [Delegate],
  [Permissions] 
FROM [dbo].[tblImport]
WHERE [Delegate] <> 'None'
GROUP BY
  [User],
  [Delegate],
  [Permissions]
HAVING COUNT(*) > 1;

I tried to re-create this using Power BI.  I went into Power Query mode, selected the three columns, clicked Transform|Group By, and created a new column called DataCount that uses the "Count Rows" Operation. 

However, the new column has the value "1" for every row, even the ones that should be duplicates with a count > 1.  What am I doing wrong?

[MFelix]

Hi @imani_tech,

To what I can see from your SQL when you do a group by the 3 fields in your table and choose the ones that have a double count you are returning a single value for row and not the duplicated values.

To replicate this though in Power BI and return the ones that are above 1  count you need to take the group by part of your SQL and then use the Group by in Power query to return 1,2, 3... in the DataCount column.

Don't know if this is just a test of the capabilities of PBI but if your data source is already refined no need to go back and ungroup and then group it on PBI, let the SQL do that for you.

Regards,

MFelix

[imani_tech]

Thanks for the response.  I'm actually supposed to complete this project within Power BI, since the data source is a .csv file.  However, I imported the .csv into SQL Server just to confirm that I'm doing the right thing.  

[MFelix]

If you upload the CSV file to PBI and do what you were doing you should get the correct result.

Regards,

MFelix

[imani_tech]

It should work but it isn't.

[MFelix · Accepted solution]

 Hi @imani_tech,

Made a small CSV file with the same fields as you have:

UserDelegatePermissions

A	AA	AAA
A	AA	BBB
A	AA	BBB
B	BB	AAA
C	CC	AAA
C	CC	BBB
C	CC	CCC
C	CC	CCC

As you can see I have repeated items, then made the grouping in Query editor and the result gives the duplicate in the count rows:

Are you already getting the information from the CSV or from the SQL that you had in your previous post?

Regards,

MFelix

[imani_tech]

I created a new query and followed your steps again.  It worked!  Thank you such much.