When I’m doing a SQL Server health check, one of the things I like showing clients is a list of their top tables by size.
Inevitably, there are a bunch of tables that:
Tables whose names start with “temp” or “backup” or “just_in_case” Tables with no clustered index that were clearly created with a select-into, and then forgotten about Tables that haven’t been accessed in months or years
We have some chuckles, make a quick calculation, and realize that our database is 10%-20% larger than it needs to be. That means our backups are taking 10-20% longer, corruption checking, restore