When maintaining workflows, for some nodes it is always possible to see the configuration regardless of node state, but for others the configuration is only available if the node immediately prior to it has been executed.
The reason for this is generally obvious, in that for some nodes it is not possible to configure it until KNIME can determine that certain data types are available to it, or there may be other pre-requisites that must be met before it can be configured. Other nodes are less “picky”
However… this behaviour can make it difficult to maintain or fix problems with a workflow.
For example, I have had situations (I’m sure it is common) where I have been making changes upstream and realised that a change I was having to make was going to directly affect node X downstream, but I couldn’t remember exactly how that node was configured and so could not determine without running the flow again exactly how I should make this current change. It would have been nice to be able to go to node X, and see how it was last configured. That’s a trivial example. I am sure that many who have written complex flows can think of better real-world examples.
Another use case that I have recently encountered was where I wrote a complex flow involved workflow that read json from an API and used the various values to populate local database tables. That was a few months back, but I now want to see how I coded/configured certain things because I want to do something similar in a new workflow. Unfortunately the old rest API has been “switched off” and is no longer available. The flow was not saved in an executed state. I therefore cannot execute the old flow, and so all I have is a “pretty piece of node-art” with no ability to inspect the workings through the KNIME 4.7 GUI, because I cannot open the configuration on the now “dead” nodes.
What would be really nice in these situations would be the option to click on a node and inspect its “last known configuration” in read-only mode. At the moment, for the latter case (where I can no longer execute the workflow), my only option appears to be to try to go behind the scene and directly inspect the xml of specific nodes to see if I can decipher the configuration, which isn’t very pretty, and is very time consuming.
What do you do in these situations? Maybe I should be able to remember how the nodes are configured, but at my age, I don’t even remember what I did at the weekend!