Find max value and use corresponding/adjacent cell?

KNIME Analytics Platform
1

Hi, I've searched for a solution on this forum but haven't been able to find anything. Thanks in advance if you're able to lend a hand =)

I'm trying to find the max value from Col1, Col2, and Col3. Once found, I want its corresponding/adjacent cell to be outputted.

 

Col1      Col2     Col3      Col1-ADJ     Col2-ADJ       Col3-ADJ       Output
1960     1966    1968            B                    M                   P                  P
1966     1970       ?               B                    D                    ?                  D
1966     1968    1975           C                    M                    P                  P
1982     1972    1969           C                    P                    L                   C
1967     1969       ?               ?                    ?                    ?                   ?

2

You could do this over a couple of steps.

first use column aggregator node. Select columns a, b, and c to aggregate by, and choose to take the maximum value. Give the output column a name, i.e result.

now connect a rule engine node. Set up three rules in there;

if col1 = result column, then outcome should be col1-ADJ, making sure is a column reference is selected. repeat same procedure for col2 and col3.

give the output column a name, I.e. "output".

 

hope this helps, Simon.

3

Simon, this worked like a charm. Thanks so much!

4

Hi,

is there a way to do this over a collection of 220 column pairs (instead of the examples three column pairs)?

For example I tried to use the column list loop start node. But doing so, I can only include one column name iterating instead of the necessary two.

Thanks,

Manu

5

Hi Manu,

You can use the Java Edit Variable node to create the second column name as a flow variable (I assume it is related to the first in some way). Otherwise, you might use the Java Snippet node too.

Cheers, gabor