For the count to ignore 0 :

Select sum(case when SomeCol = 0 then 0 else 1 end) as NewCount from dbo.YourTable

It's hard to create the actual query without the...

You can join a view and a table. You can also select from another database using 3 parts name >>

Select A.col1, B.col2 from dbo.TableA A inner join MasterDb.dbo.ViewName B...

I remodified the code (without doing the param thing Noeld proposed).

modifications :

--this clears the cache and reinsert the cacheable plans.

MyCn.Execute ("DBCC DROPCLEANBUFFERS")

...

Do I still do nothing in the proc (since it takes a plan now)?

Can you do it on a pc that has the server on the same machine?? Would give out much better numbers :

the proc :

Create proc dbo.Test

as

go

the vba code...

As long as it works...

This seems to be the most consistant case (except for 100 loops.. should be soemwhat lower than that).

Loops : 1       RPC : 0     SQLBat : 0        0 %

Loops : 10     ...

There are too many test cases. Here I'm interested in the difference between RPC and SqlBatch. The execution time of the sp is always gonna vary depending on...

Alright... good luck with the rest of the project.

Don't do that, the more precise and accurate info we have, the better answer we can provide.

Now you can tell him he wasted a few dimes for absolutely no reason and that you could have told him that before he bought it

Well you could always insert the results in a temp table, extract the count from there and reselect the results... but that obviously is somewhat slower. Now it depends...

Actually I'm thinking that my test doesn't really test the different between RPC and SQL batch... I'll remove the select in the proc and the parameter and try again.

Momentary ambition. Not interested in its job... Too much work (70 hours/week) .

Use the local parameter settings on the clients... That'll present the data as they wish (by itself if using excel of access). Simpler and makes less work for the...