This would be a constantly used feature by every user. The inability to sort off of the generated Row IDs is very frustrating. The KNIME concept of “Row ID” seems to be based on Joining data and maintaining permanent IDs throughout the entire workflow. However, users often have to use what I call temporary numeric “Process IDs” which are used to track rows, sort and error check when going through heavy transformation processing and do not necessarily persist through an entire workflow. A simple setting for numeric row ID would be a much cleaner solution than using the Math Formula node with “ROWINDEX+1”.
thank you for your suggestion. I have created a ticket for this (internal reference: AP-17720).
Instead of the Math Formula node, you could also use the Counter Generation node. Does the same as the ROWINDEX+1 function but might be more intuitive.
I will look into the counter generation node in the mean time.
Hi @iCFO , the Counter Generation node is an excellent option for that. It’s what I use all the time for what you described.
I almost never rely on Knime’s RowID if I want to really identify a row after order is changed. I always create my own ID column using the Counter Generation.