Domain Aggregate Functions

  • Thread starter Thread starter Guest
  • Start date Start date
G

Guest

Hello,

I have been trying to migrate my Access Application built in Access 2003
onto SQL Server, but for many of the forms I use Domain Aggregate Functions
(DAFs). I am still keeping my front end as Access but the back end process is
SQL 2000. I decided to use ODBC linked tables instead of an Access Project as
it is easier to upsize the backend (to SQL Server 2005). At the moment the
volume of data is reasonably small, but it is expected to increase
expotentially, so I need to minimise potential performance issues.

Does anybody know what the inner workings of these functions with ODBC
tables...do they do a full table scan before executing a select, or another
aggregate query, or is the processing done on the server side?

Do DAFs form an independent connection to the linked table each time you
call them?

Is using an ADO connection object faster and less resource intensive than
DAFs?
 
Hi,

I have tried using pass through queries but they execute much more slowly
than the DAFs on ODBC linked tables. Does the DAF use the DBEngine to process
information from ODBC liniked tables or is it using ODBC Direct connection?

--
Grassy,
Try it you might like it :)


Alex Dybenko said:
Hi,
in general DAF works ok, but good idea - to build your own DAF functions,
which will run pass-through queries, so you will be sure that query executed
on a server

--
Best regards,
___________
Alex Dybenko (MVP)
http://alexdyb.blogspot.com
http://www.PointLtd.com
Click to expand...
 
hi Grassy,
I am still keeping my front end as Access but the back end process is
SQL 2000.
Click to expand...
Why not using SQL 2005? The usage of schemas makes develpment much more
transparent as you can save your form recordsources under Form or report
sources under Report.
I decided to use ODBC linked tables instead of an Access Project as
it is easier to upsize the backend (to SQL Server 2005). At the moment the
volume of data is reasonably small, but it is expected to increase
expotentially, so I need to minimise potential performance issues.
Click to expand...
This is imho the easiest way, but you have to keep in mind:

Let the server do the work. So normally you need a lot of views - more
than queries before.
Does anybody know what the inner workings of these functions with ODBC
tables...do they do a full table scan before executing a select, or another
aggregate query, or is the processing done on the server side?
Click to expand...
The built-in DAFs may kill performance, you may use Michael Kaplan's
replcement functions instead. But remeber, often it is bette to let the
server gather the information.


mfG
--> stefan <--
 
Grassy03 said:
Hi,

I have tried using pass through queries but they execute much more
slowly than the DAFs on ODBC linked tables. Does the DAF use the
DBEngine to process information from ODBC liniked tables or is it
using ODBC Direct connection?
Click to expand...

In my experience the DAFs send a very nicely formatted SQL statement to the
server so it can be executed there. If they perform to your satisfaction then
just keep using them.
 
Back
Top