Trying to eliminate cross column duplicates

[Anonymous]

Hi,

I am building a check for my database to eliminate duplicates. Straight up duplicates are easy to remove but sometimes people make typos or something, and I need to remove those too. 

This is a rough example of my data:

Reference	Date	Description
ref1	1-1-2020	Blub
ref2	3-1-2020	Blob
ref3	4-1-2020	Blab
ref3-1	4-1-2020	Blab
ref1	1-1-2020	Blib

Now I fuzzy matched it on itself to create this: 

Reference	Date	Description	Dup.Reference	Dup.Date	Dup.Description
ref1	1-1-2020	Blub	ref1	1-1-2020	Blib
ref2	3-1-2020	Blob	ref2	3-1-2020	Blob
ref3	4-1-2020	Blab	ref3-1	4-1-2020	Blab
ref3-1	4-1-2020	Blab	ref3	4-1-2020	Blab
ref1	1-1-2020	Blib	ref1	1-1-2020	Blub

As you can see the same record effectively is still double, appearing once on the left as x-y and once on the right as y-x.
Now I need to find a way to eliminate these cross column duplicates and end up with this:

Reference	Date	Description	Dup.Reference	Dup.Date	Dup.Description
ref1	1-1-2020	Blub	ref1	1-1-2020	Blib
ref3	4-1-2020	Blab	ref3-1	4-1-2020	Blab

I need both sides to check whether the double record really is double or intended that way. 

Thank you!

[Greg_Deckler]

Power Query or DAX?

In DAX you can remove duplicates by using SELECTCOLUMNS to select each column or columns into a table VAR. Then UNION and then DISTINCT.

[Anonymous]

Hej,

No M in Power Query.

And I'm not sure how that would work... 

Thanks though 🙂