Reference Table: Are filtered-out rows still there?

[Anonymous]

Hi, I've created a reference table from a table which has no unique key. 

In the reference copy, I filtered out all the duplicates (by Grouping the column) and am left with only unique rows, which I thought I was going to be able to use as a unique key. 

But every time I try to join this table to another table, it tells me it's a many-to-many relationship. I'm assuming that the filtered-out rows are still somehow present and that this is why it thinks nothing is unique: even though I can see only the unique rows, Power BI is referencing the original table, seeing all the rows. 

Hoping someone can confirm that this is expected behavior. I think it is, and want to be able to stop looking for other ways around it. 

[jbwtp]

Hi @Anonymous,

if you referenced a table in Power Query Editor and then manipulated with this new query, it should not default or reset to the original query by any means, including restating rows (or even manifesting that they were ever existing). From PBI Designer point of view, these are now two completely separate tables on its own rights.

There amy be something else happening in this case.

Could you please check that you are using a correct table for the unique key and the unique key table does not have bi-directional relationships with the original query, better remove any relationships between them while testing.

Cheers,

John