I was wondering if it was possible to continue past a failed node or a node that hasn’t ran. I am trying to create a level of error handling on my process, in which if the process runs I can have the node at the end, pass a variable to a sharepoint connector to get it to run and write into an excel “process ran” and then it would insert the date.
However, if the process fails at some point before that, i would still like the sharepoint connector to run and the excel document be written in but with “process failed”.
I believe the difficulty is because it is because the node hasn’t “failed” and instead is just finishing with errors. Therefore the “catch” wont complete. Is there a way to get around this?
The node is not executed to be failed with some error that the try-catch costruct can handle. It’s a configuration error and you need to configure the node properly to avoid it.
For example, this can happen if the node is configured by a particular flow variable which is not available anymore.
How do i configure the nodes to fail fully if they dont execute? Basically this is running on our KNIME server and i was wanting to create some form of error log. So if it runs it will write into an excel on a sharepoint that says “success” but if it fails, i want it to note the date and time it failed.
Oh okay, i wasn’t aware of the breakpoint node. I dont have an “error message” exactly, im trying to get it to fail on purpose.
Basically what i want to happen is, if any of the nodes dont execute for any reason, incorrect data, missing data, database is down…etc. Where ever it breaks, i wanted it to “Fail” so then the node could pick up that it has failed, that would finish executing but it would write into the excel “process has failed”
You can handle failed executions (runtime errors) by using the try-catch construct. But you have to avoid configuration errors. In the screenshot you shared, the error is a configuration error, so you need to check what is missing in the configuraitons and provide a solution to prevent this from happening.
For example, if there is a missing variable (a variable that you expect to exist and used it in the configurations), the node will not execute with a “configuration error”. So you can (automatically) check if the variable exists before the execution of the node and if it’s missing, fail the flow by using the Breakpoint node.
