Sorry for posting so much, but since I am considering using KNIME as integral part of my PHD-work I am testing some basic stuff.
I noticed, when using a Python Node (Source; 1=>1) creating a dataframe that has no columns but an index, after node execution that index is lost. Knime is able to handle that situation in other circumstances. You can perfectly use column separator and separate all columns having just the Index left. You can even feed that column-missing data into a python node which will have a input table containing the correct index. Just behind the python-node the index is basically gone.
I also appended the workflow which illustrates that behaviour. (Inspect the output tables)
Bug_flow.zip (27.8 KB)