with all the problems of the

.ipynb

format, it looks like it refuses to die. just got this on Databricks

I mean this is better than their weird proprietary standard until now

interesting - does it has a different look? would

%load_node

now works with it?

nope it does not, in fact it does a very weird thing

I don't think this is testing the right thing. Currently the logic is we auto-detect if it's on databricks, if so it will print the code instead of creating a cell due to the previous limitation. To properly test this, you need to remove that hardcoded logic, or maybe mock that function we used for checking (something like

_is_databricks

, don't remember exactly). In addition, the weird result here could be how the "print" works differently with this"new notebook", IIRC it's

rich

based so maybe something to look into.

makes sense. happy to let someone else try it 😅 or I will do it in a few weeks

I did a quick test and it doesn't work with

ipynb

(it's the default). I test it with the IPython

%load

command and it doesn't create any new cell even run successfully. Kedro use a similar logic so I don't it will work.