Thank you for your feedback.
This was a bug, we fixed it and fix will arrive at production very soon. Please, keep us updated if you will experience any problem with code evaluation. We will create a post with detailed explanations of how our incremental kernel works, why such problems could appear and what you can do to avoid them.
Hi, I checked it and it’s definitely a bug of a Datalore Kernel, it’s experiencing troubles with finalizing and saving cell state after execution when there is open connections left. We will fix it and while we are working on it you can still workaround it explicitly closing connection in the same cell you created it.
I would suggest to use something like:
with engine.connect() as connection:
# Do some loading data stuff here