Bug in GroupBy Node - Duplicate Columns

Hello,

 

It's difficult to explain this issue without referring to the KNIME workflow I've attached but here's what I'm observing:

 

With the Data Generator node (though it could just as well be any set of Double and Integer columns), using Type-based aggregation of both Double and Int types produces duplicate output columns (in this case, I get 2 Lists output for each input Integer column).

 

Can anyone else confirm this?  Again, see the attached workflow.

 

Ed.

Hi Ed,

this is not a bug but the expected behaviour. The integer columns are added twice since they are also DoubleCell compatible. So if you want to apply an operation on all numeric columns you simply have to select the DoubleCell but not the IntCell, LongCell and DoubleCell. Integer columns are also compatible with LongCell but not the other way around. So in your case you would only need to add the DoubleCell in order to get a list representation for all numeric columns.

The reason for doing it this way is that KNIME can also apply an aggregation to a numeric column type it does not know as long as the unknown type is compatible to DoubleCell.

Bye,

Tobias

Tobias,

 

Ok thank you for clarifying that.  I don't recall seeing that in the node's instructions but I didn't read it that hard :/

 

Ed.

Hi Ed,

in the node description I only describe that you can use DoubleCell to aggregate all nuermic columns but didn't explained it that detailed. I will update the description accordingly for the next release. Thanks for the hint.

Bye,

Tobias

Alright, thanks for the help!

 

Ed.