Well you could always store a dictionary (or other Python object) as a pickle file and later reload it. Not sure if this is of any help.
The Python dictionary could be converted into a proper KNIME table. And later back to a Python dictionary. The exact settings and structure are a little tricky - you might have to toy around with it.
Bring back the dictionary from the Dataframe and from the pickle file and compare them (links to some relevant Stackoverflow entries in the file)
kn_example_python_pickle_dictionary.knwf (22.5 KB)