A far better way, IMHO, is to write a good SELECT in a stored procedure. 😉

What's the old saying? "Just because you can do it in Python, doesn't mean you...

It's coming... According to the way they handled things like GENERATE_SERIES(),...

@steve-2 Collins,

Sorry... I tried your latest code on a max value of  "just" 1 million.  I stopped it at 2 hours and 8 minutes.

It's coming... According to the way they handled things like GENERATE_SERIES(), etc, it should...

Ok... I'm a little gobsmacked... I ran your code with zero changes and it's riddled with syntax errors and,yet, it ran without error.

What is the following supposed to...

I also have to admit that I have no clue what something like this is supposed to be doing in T-SQL.  In another language, maybe but not in t-SQL.

(prime <<...

Ok... I'm a little gobsmacked... I ran your code with zero changes and it's riddled with syntax errors and,yet, it ran without error.

What is the following supposed to do for...

So why are you using...

The doubled-up relational symbology isn't floating my boat on an SQL Server forum, either.

Sounds good.  You could still be affected if you copy a DATETIME2(7) into any datatype that has a lower resolution.

And, for a change, I shall be...

I tried your script with 10 million rows and it took 25 seconds with dbo.fnTally using 126 MB log file. With GENERATE_SERIES it took  it took 126 seconds using...

Heh... this reminds me of every bloody release since and including 2005.  The ones before that didn't seem nearly as bad.  Maybe because there were fewer moving parts.

The spooky part...

Heh... or resolve yourself to the fact that no one will make any changes to the code no matter what you find and just let them do the testing they're...

From my testing SQL Server's GENERATE_SERIES is efficient and well-optimized. It's not surprising it performers better than user defined functions as it is implemented in the native code of...