I am trying to identify errors coming from Databricks. So I can handle them in my code.
Sometimes I get a descriptive error, that points me to the exact problem, but then if I run the exact same test, I sometimes get "request error: invalid operation state. This should not have happened" .
If there was some persistency I could work with that, but it looks like am getting different result with each run.
What might be the reason for that "request error: invalid operation state. This should not have happened" message and how I can avoid it, so I will always get a proper error message from Databricks.