Suggestion to consider

[BondyB]

Hi Andy,

Love what you have done here for generating MLVs and creating a repeatable deployable code pattern.

Just a thought and something you may want to consider if / how to handle in the flow?

Scenario

A user deletes an MLV manually (or via code) in the lakehouse.
However, the MLV SQL definition file still exists in the lakehouse file but is unchanged since the MLV was originally created.
Since the metadata file still references the file and the file is unchanged it does not recreate the now deleted "missing" MLV(s).

There is a question here as to what the right decision in the flow would be right?
Should it recreate the missing MLV (Which someone purposefully or accidently deleted)
Or should it flag that the MLV was deleted but the SQL file still exists.
Or should it archive the MLV SQL definition file.

Be interested to hear your thoughts on this one

Cheers

[BondyB]

Further to this -
If the metadata file is "deleted" - this would create the MLV that was deleted if the SQL definition file still exists.
Certainly some thoughts around handling these kinds of scenarios