Extract hh AM/PM from getdate()

Hi,
I am looking for a query to extract hour and AM or PM value from a date on sql2000.

ex/-
Input : 2001-12-28 22:18:07.810 (from getdate())
Output : 10 PM

select convert(varchar, (datepart(hh, convert(varchar, getdate(), 8)) % 12)) + ' ' +
substring (convert(varchar, convert(datetime, getdate(),20), 100),
DATALENGTH(convert(varchar, convert(datetime, getdate(),20), 100)) - 1,
DATALENGTH(convert(varchar, convert(datetime, getdate(),20), 100 )))

The above works but is there a better way to do this?This is a little shorter:

SELECT CONVERT(VARCHAR,DATEPART(hh,GETDATE())%12) +
CASE WHEN (DATEPART(hh,GETDATE())%12) > 0 THEN ' PM' ELSE ' AM' END|||thanks for your reply.
but i figured that 12 AM or 12 PM was displayed as 0 AM and 0 PM.
Hence to reduce my troubles, i will stick with the good ol' substring.

SELECT (substring(CONVERT(VARCHAR,getdate(),22),10,2) + ' ' +
substring(CONVERT(VARCHAR,getdate(),22), 19,2))

External stored procedure, performance?

Hi everybody,
I should begin to write a DLL library for Sql2000 server.
The functions I like to implement are mathematical functions, like standard
deviation, and similar, nothing really complex. Often I have to use more
then one standard deviation inside the sama function, using subset of a
record set.
Of course I can use sql2000, that implemets standard deviation and basic
mathematical function, so the question is: is it opportune to write a DDL to
improve performance, or is it worse, or just the same?
thanks a lot for any kindly advice
cesareI think the key words here are "using a subset of a recordset". Based on
that I would implement it in T-SQL, or on the application side but not in an
XP. The cost of connecting back to the SQL Server to grab a subset of rows
is going to be big, especially if you perform this std dev calc several
times in a row.
"Cesare" <cvairetti@.mcgestioni.it> wrote in message
news:uO9XdpvjGHA.1640@.TK2MSFTNGP02.phx.gbl...
> Hi everybody,
> I should begin to write a DLL library for Sql2000 server.
> The functions I like to implement are mathematical functions, like
> standard
> deviation, and similar, nothing really complex. Often I have to use more
> then one standard deviation inside the sama function, using subset of a
> record set.
> Of course I can use sql2000, that implemets standard deviation and basic
> mathematical function, so the question is: is it opportune to write a DDL
> to
> improve performance, or is it worse, or just the same?
> thanks a lot for any kindly advice
> cesare
>sql

Extended stored procedure, performance?

Hi everybody,
I should begin to write a DLL library for Sql2000 server.
The functions I like to implement are mathematical functions, like standard
deviation, and similar, nothing really complex. Often I have to use more
then one standard deviation inside the sama function, using subset of a
record set.
Of course I can use sql2000, that implemets standard deviation and basic
mathematical function, so the question is: is it opportune to write a DDL to
improve performance, or is it worse, or just the same?
thanks a lot for any kindly advice
cesare"Cesare" <cvairetti@.mcgestioni.it> wrote in message
news:et2MIqvjGHA.4304@.TK2MSFTNGP03.phx.gbl...
> Hi everybody,
> I should begin to write a DLL library for Sql2000 server.
> The functions I like to implement are mathematical functions, like
> standard
> deviation, and similar, nothing really complex. Often I have to use more
> then one standard deviation inside the sama function, using subset of a
> record set.
> Of course I can use sql2000, that implemets standard deviation and basic
> mathematical function, so the question is: is it opportune to write a DDL
> to
> improve performance, or is it worse, or just the same?
>
Extended stored procedures are so dangerous to the stability of the database
server that they should be used very, very carefully.
In SQL 2005 CLR integration provides a safe way to extent the SQL engine
with custom calculations.
David