The new inline script feature in Fabric notebooks is great. Inline sits right after data cleansing, just before the data frame gets persisted (append or overwrite, as applicable).
But sometimes we need scripts that are executed either before table notebook execution starts or after it has finished. An example of the latter is when something needs to be done on an entire incremental table and not just the delta data frame.
This can be achieved for example by dummy tables with inline scripts and a lot of manual object dependencies (to get the execution order right), but that’s an inelegant solution with a high risk of missing required object dependencies.
Furthermore, dummy tables count towards the painful Fabric workspace limit of 1000 items.
It would be good if we, in addition to the current inline scripts, could also have pre- and post-scripts as wrappers of the main notebook. This would handle object dependencies / execution order and we wouldn’t have dummy assets that count towards the 1000 items limit.
The pre- and post-scripts should become part of the main notebook even though the developer most likely won’t benefit much from any notebook context (as opposed to the current inline script feature).
