Hi!
I have a suggestion, not a platform error.
We have been using KNIME quite a bit over the last year,
and I have noticed that when an input (an Excel file, for instance) is changed, and a variable that was used later in the processing no longer exists,
the node that uses that variable randomly modifies it to another one (for example, in a rule engine).
In the interest of providing stability to a processing workflow,
it might be better if the node did not change and instead indicated as an error that the variable no longer exists.
Otherwise, it becomes complex to know where my original flow has changed,
which destabilizes the calculation.
This is especially true when adjustments are made to the input data
or re-adjustments to initial nodes.
Thanks in advance.
Best regards,
L.