I use Azure DevOps git repos for all object definitions within a schema.  This includes table and view definitions, but also things such as column metadata, ownership, and permissioning as required.  These same notebooks are dynamic in nature to allow us to promote schema and object definitions to higher environments through CI\CD pipelines.  As changes are required to tables, we increment version numbers so we can apply them in order and develop them to be re-runnable.  For example:

ServiceNow Object Definitions - Release 1.0 - Contains the initial table and view creations.

ServiceNow Object Definitions - Release 1.1 - If we decided to include new columns on the incident table this script would contain the ALTER TABLE ADD COLUMN statements required.

Now for promotion to Test or Production, we can decide to run all versions (assumes empty environment) or just 1.1 if 1.0 has already been deployed.