For comparison purposes the date and time columns could be combined in a virtual table ('conv') into datetime columns.  Then maybe 3 separate CASE statements so as not to overlap...

• This reply was modified 5 years ago by Steve Collins. Reason: Fixed error in code

He didn't check with you first before writing the article?  Is it because money and smallmoney are really bigint and int?  I'm pretty sure major payment API's like Stripe uses...

Here's one by Phil Factor on red-gate.com

Avoid use of the MONEY and SMALLMONEY datatypes (BP022)

What about an open source SQL Server project to create a generic forum with no dependencies on UI and which allows for a 3rd party OAuth 2 identity provider?  The...

Imo a type conversion to an approximate data type and a comparison to 0 are not strictly what the question asked.  As you know well float and money data types...

All right Sergiy's correct about the CROSS APPLY's not being necessary.  What I don't agree with is using MONEY (VARCHAR converts to FLOAT too fwiw) and comparing the sum to...

On second thought, instead of individually calculating the decimal costs they could be added all together into one value 'dCost'.  Also, a combination of TRY_CAST and ISNULL could be used...

The obligatory comment is regarding the problems arising from storing numeric data as varchar instead of decimal(#, #).  Assuming you're stuck with the DDL you could try something like this. ...

It appears you're looking for $137,000 less the running total of yearly license costs.  Also, the 'money' data type is imprecise and has serious issues so this uses decimal(14, 2)...

Simplified using OUTER APPLY instead of the LEFT JOIN

with
max_min_cte(t_min_dt, t_max_dt) as (
    select min(dateadd(hour, datediff(hour, 0, start_dt), 0)),
      ...

This uses a tally function (or it could be called a "numbers function") twice.  Once, to generate the total hourly datetime range across all rows in table #t.  Second, to...

select *, dateadd(hour, datediff(hour, 0, in_datetime), 0) out_datetime
from #data;

[EDIT] Same answer as Ant-Green but 2 minutes later 🙂

select left(isnull(stuff(first_name, 11, 3, '...'), first_name), 13) 
from @t1;

Another alternative could be CROSS APPLY

update t
set AmountAll = ca.sum_amount
from #test t
     cross apply (select sum(tt.Amount)
       ...

As long as people are sharing their CSV secrets.  Warning disclosure this contains C# code.  A couple of years ago I worked on a project for a fitness tracking system...