Based on the limited info we have so far, the JOINs would look something like this:

SELECT *
FROM dbo.main_table mt
LEFT OUTER JOIN dbo.other_table1 ot1 ON ot1.WOQty = mt.WONum
LEFT...

DELETE FROM dbo.Table WHERE ID IN (12);

...

With the temporal table, deadlocks are guaranteed every time. Only one of the deletions will succeed; the rest will fail.

We have tried ......

Many index-suggestions I get from the database I'm working on, very often suggests quite alot of include-columns. Could this be a sign of poor database-design?  As far as I...

No idea.  They're not system tables that are intrinsic to MS.  Some other user/3rd party code/script must be creating them.

Heh... more...

BWAAAA-HAAAA-HAAAA!!!  Ok... I "get it".  Yep... I agree... you never have to mention "IDENTITY" in a logical model.  I have to tell you, though, that the idea of having...

But, IIRC, the default trace does not record what id specifically made the change if that id has certain levels of permission.  I think if you want to know the...

A DDL trigger should be enough, including putting in place things to monitor it, to make sure no one has modified that trigger.

In a business environment, no one should be...

As a first step, make sure you are on the latest CU patch level.

That's where you and I would differ... to me, the "customer number" would be the IDENTITY column.  Or, perhaps, a Random GUID. 😀  Why would your "customer...

It would be really interesting to see you design a Customer or Employee table without one.  I'd also be interested in your claim of "it's literally impossible to normalize...

Heh... so THAT's the reason why so many people end up with "slow databases" and IDENTITY columns everywhere.  😉  A lot of the important stuff isn't even considered up...

Great, sounds good.

Again, though, your initial design should NOT be "tables", it should just be data.  It should be at a business level, not at a techy level.

If at all...

Then definitely start the design before tables.  That is, do a logical data...

[Use ROW_NUMBER() partly because of inherent efficiency] And partly because it's so easy to adapt:

SELECT theDate, value, flag
FROM (
    SELECT *, ROW_NUMBER() OVER(PARTITION BY...