If you have FK relationships, you don't have to do the data changes yourself.

Script out the FKs, "DROP and CREATE", and change the "ON UPDATE CASCADE" to "ON...

I don't know about detach at this point, but with some "fancy footwork" you can DROP a database only if it was OFFLINE.  You have to rollback the DROP, of...

Or perhaps multiply by 2.2046226218, or some shorter variant of that, if you need more accuracy.  2.2 is rather rough estimate.

Maybe easy, maybe not.  You have to create a synonym for every remote object you reference.  And keep them in sync if, for any reason, you have to rename an object.

That's going to be a royal pita.

Why not just stop using different names for dev and prod?  It's easier to mess up the aliasing than it is to...

From Books Online, under sys.dm_db_index_operational_stats:

range_scan_count  bigint  Cumulative count of range and table scans started on the index or heap.

Go by the values in the clustered index.  For example, if the table is clustered on a datetime column, use that as the value for the UPDATE.

Here's a...

The key lookups should cost more than the sort for only 7 rows (I know the lookup is in both queries).  You should consider adding DateRegistered to the NC lindex.

Be sure to create a unique clustered key / primary clustered key on SiteKey in the #Sites tables.  With only ~20+ rows it may not matter, but it can't hurt.

DayOfWeek 7 is ambiguous in SQL Server, since SQL allows the week to start on any day that you specify.

Instead, I'd strongly urge you to use known day...

If you can, TRUNCATE the table rather than DELETE all the rows, as trunc will be much less overhead (btw, in case you're wondering, a trunc can be rolled back...

Sorry about the HOUR(), that's the one function I always forget doesn't exist.  Per quarter for the current year, something like this:

select
  DATEPART(quarter,...

Not 100% sure what you need, but, for example, this gives sales for the latest quarter for >=7AM and <8AM:

select *
from TicketItem
where s_item...