As an alternative to a CASE expression, you can use the SIGN function:

select column1, SIGN(column1) * POWER (abs(column1), 1.8) * CAST(10000 as float) from #t

It runs without any error, but what does it mean? Second, CAST(10000 as float) is equivalent to 10000. (with a dot).

Peter

Thanks Lynn, it's a long time ago I opened a math book but that's what I remembered from it. I'm curious what the OP tries to compute.

Peter

I don't know what you are trying to compute, but there's no workaround for this simply because it's mathematically undefined. You can't raise a negative number to power which is...

Glad we could help. Thanks for the feedback.

Peter

Chris' solution returns 9 rows because it only take 3 date points into account, the same as your original query. My solution returns 12 rows because it take 138 (3...

Do you mean you have 140 almost identical query's UNIONed together? Try this one:

SELECT

Registration_ID, [Person_ID], Start_Date_Registration,

J.Duration, A.[Date Point]

FROM

dbo.Test_Registrations

CROSS JOIN

(

SELECT...

Here's an example with @i <= 65536

declare @i int = 65536;

with

Tally4(N) as (select 1 union all select 1 union all select 1 union all select 1) ,

...

Yep, actually it goes wrong when @i > 2048. spt_values contains numbers from 0 to 2047. If @i can be greater than that in your situation then you have to...

Apart from the order of the result, I don't see what's wrong with your own solution. If you want the result in the order you specified in your example try...

Peter Brinkhaus (3/4/2010)

---

As an alternative for those not using SQL Server 2008 SP1 CU5 or working with SQL Server 2005, you can duplicate the query each with a different condition...

As an alternative for those not using SQL Server 2008 SP1 CU5 or working with SQL Server 2005, you can duplicate the query each with a different condition and UNION...

Maybe this one?

SELECT

CASE

WHEN A.K IS NULL AND EXISTS (SELECT * FROM A WHERE A.K = B.K) THEN

...

I have performed the test too with similar results. All solutions benefit from an index on columns Section and DT. Paul's solution benefits the most from the index and becomes...

Be careful. Depending on the collation being used you can get very odd results when comparing characters. For instance, I am using collation Latin1_General_CI_AS and the insertions into table Replace_Lookup...