Remove symmetrical records with Power Query

[webportal]

Is there a way to remove the yellow records with Power Query?

They are errors and I need to get rid of them:

The process could be something like:

1. Filter the table to retrieve the rows containing only the current "CustomerID"
2. Check the Sales. Is there any value with -Sales?

And then delete those rows.

This table contains many other Customer ID's - this is just a sample.

[v-yalanwu-msft · Accepted solution]

Hi @webportal  ,

According to you description, you could Use M language in power query, as follows:

let

#"Changed Type" = Table.TransformColumnTypes(Source,{{"CustomerID", Int64.Type}, {"Sales", Int64.Type}}),
#"Added Custom1" = Table.AddColumn(#"Changed Type", "Original", each Number.ToText([CustomerID]) &" "& Number.ToText( [Sales])),
#"Added Custom" = Table.AddColumn(#"Added Custom1", "Opposite", each Number.ToText([CustomerID]) &" "& Number.ToText( [Sales]*-1)),
#"Added Custom2" = Table.AddColumn(#"Added Custom", "Flag", each if List.Contains(#"Added Custom"[Original],[Opposite]) then "" else 1),
#"Filtered Rows" = Table.SelectRows(#"Added Custom2", each [Flag] <> null and [Flag] <> ""),
#"Removed Columns" = Table.RemoveColumns(#"Filtered Rows",{"Original", "Opposite", "Flag"})
in
#"Removed Columns"

The final output is shown below:

vs

Best Regards,
Community Support Team_ Yalan Wu
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
#End Reply 2nd

[Ashish_Mathur]

Hi,

Your question is not clear.  Why should the rows showing 222.83 and -222.83 not be deleted as well?

[webportal]

@Ashish_Mathur 

Hi, yes. Those rows should also be deleted since they're simmetrical.

[Ashish_Mathur]

Sorry but i won't be able to help with this.

[Vera_33]

Hi @webportal 

If I understand correctly, you also need to remove the two rows with 222,83 and -222,83 for this Customer ID in the sample? And the rows you want to remove are always together? Can you provide some sample data in a format we can copy?

[webportal]

Exactly @Vera_33 

No, the rows aren't always together, they can be far away.

[Vera_33]

Hi @webportal 

I am asking do they appear as a pair like 222,83 and -222,83, or 222,83 can be far away with -222,83?

[webportal]

No, they don't always appear as a pair.

Sometimes they're far away.

[Vera_33]

ok, @webportal  have you tried the way from @v-yalanwu-msft ?

I've got anther 2 questions:

#1 does postive always come first in the pair? say  (222,83 and -222,83), 222,83 comes first?

#2 is it possible you have 222,83 and -222,83 and 222,83 for the same customer or 222,83 and 222,83 and -222,83?

Really need to observe the data and find out patterns