That's Steve Jones.  He's the...

As per Phil needs data.  Hmm... either the rows would need to be expanded or dynamic sql?  In order to use ROWS BETWEEN N PRECEDING we're stuck with N being

Sorry I shouldn't have wrote that.  I love horses and racing too.  Yes, it passes my visual test now 🙂

After looking at it again (and again) it was updated a few times

Anyway code could be inferred from what's there.  I don't want to know about horse racing.  Untested you could try

drop FUNCTION [dbo].[udf_GET_EACHWAY_RETURN_itvf];
go
create or ALTER FUNCTION [dbo].[udf_GET_EACHWAY_RETURN_itvf](
 ...

• This reply was modified 2 years, 8 months ago by Steve Collins.
• This reply was modified 2 years, 8 months ago by Steve Collins.
• This reply was modified 2 years, 8 months ago by Steve Collins.

Oops posted in the wrong spot

• This reply was modified 2 years, 8 months ago by Steve Collins.

It not a fully asked question imo.  It's a good one tho.  Hopefully some of that medication was prescribed.  You're looking for more efficient code.  It comes down to demonstrating...

It hasn't been necessary for my work so...

Ha my exception is for VIEWS with INSTEAD OF triggers

Having performance issues while filtering a view is not uncommon.   Is the view indexed?  How much latitude is there for making changes?

Still works

select *
from drview
where field2 = 'x';

Works.  Not sure what the issue is

drop view if exists drview;
go
create view drview
AS
with cte as (
  select *
  from (values ('x', 'x'), ('x', 'x'), ('x',...

HNY SSC!  Thank you Steve for being a great host.  All the things 😉

I could be wrong but that looks like it will do a full scan of the table for each grouping of rows.  Why not just do a single pass...

One approach could be to use a WHILE loop to increment grp values based on row counts and the DATEDIFF interval

Starting with Jeff's code:

--=================================================================================================
--   ...