I was debugging a python code block where requests = multi_createRequest_new.data was working, but requests = pd.DataFrame(multi_createRequest_new.data) was not. I thought it was a pandas bug so I added pandas in the libraries for the workflow (since this was a fix a while ago). That seemed to make it worse. Now I see a couple of other errors:
Very odd bug, I would think that deleting and re-installing the libraries would have fixed things. Did you have numpy installed before taking the screenshot?
You said that you can't delete the libraries, do they not visually disappear when you click '...' then delete?
I will ping our workflows engineers to see if they might know what is causing this error.
In the meantime to trouble shoot, I want to suggest deleting the library, giving the workflow a practice run where it can execute successfully without the use of pandas or numpy, then re-install the latest version of each (2.2.2 for pandas).
It looks like you might have already tried that from the screen shot but my guess is that the code executor might have done something odd like cached the lib even after deleting it somehow.
My other thought to get rid of this ghost bug would be to make a new workflow from scratch and see if moving the logic over has the same bug. I tried to reproduce this but it didn't hit the same error, let me know the steps if you can reproduce the bug on a fresh workflow!
