I have a pandas DataFrame with 3 columns: col1 contains lists, col2 contains dictionaries, and col3 contains NaNs:
dict_ = {\'col1\': [[\'abc\'], [\'def\', \'ghi\