Grand Total Count incorrect after incremental update

I've been experiencing a strange behavior with a count measure where the total doesn't show the correct number after an incremental update.

Basically I have about 70 million rows in my fact table and initially I do a full process of the cube to get it up to speed. Then each subsequent day I do an incremental update to that cube adding only new rows. The strange part is my grand total count seems to do it's own thing after each incremental update.

Here's how I discovered the problem:
I made a backup of the AS database and then ran the same test 3 times restoring the original database in between each run. Each time I added about 41K rows to the cube incrementally. Each time I got a different total for the count measure. The first time it only added about 12K to the total count (instead of 41K). The second time it added about 25K. The third time it *subtracted* about 6K from the total count. How is it possible that I can add 41K rows to the cube and have the total come out smaller than it started?

When I dug into it a little more I found that if I looked at the date that the 41K rows were being imported into I saw the correct number imported into the cube. So, it's like the cube processed the rows correctly, but somehow got confused when calculating the grand total. What would cause this? Is this a bug?

How do you do the incremental update of the cube?

Do you create a new partition and insert new rows there?

Edward.
--
This posting is provided "AS IS" with no warranties, and confers no rights

|||

We are using SQL Server 2005 Standard Edition which doesn't support multiple partitions, so no, just one partition. Standard edition doesn't support proactive caching either, so we don't use that and I don't *think* caching is a problem. I went ahead and ran the aggregate designer and had it optimize the cube to 95% and then set the incremental update to look to a view that contains only the new rows that are not in the cube currently. Then when I run the update I get the random total like I described in my first post. But, like I also mentioned in the first post, if I browse the cube and drill down to the new information, all the expected fact table rows seem to exist in the cube. At the day level (when using the time dimension) all the numbers add up as they should. It's only the rollup totals that seem to be wrong.

|||

I would urge you to contact product support for analysis services and report this problem.

Edward.
--
This posting is provided "AS IS" with no warranties, and confers no rights.

GPRS replication troubles : 28562

Hello:
I get this number of errors when I'm trying to merge replicate over GPRS. It
occurs only some times:
error 80072EE7 with native 28562
error 80072EFD with native 28562
and after those errors I get 80004005 with 28557.
What about are these errors? why the database gets unstable? how can I avoid
that?
Thanks a lot
Where and how do see this errors? Can you give more context?
This posting is provided "AS IS" with no warranties, and confers no rights.
OR if you wish to include a script sample in your post please add "Use of
included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm"
"Juan" <juan@.mailinator.com> wrote in message
news:%23aF6Shg1EHA.2112@.TK2MSFTNGP15.phx.gbl...
> Hello:
> I get this number of errors when I'm trying to merge replicate over GPRS.
It
> occurs only some times:
> error 80072EE7 with native 28562
> error 80072EFD with native 28562
> and after those errors I get 80004005 with 28557.
> What about are these errors? why the database gets unstable? how can I
avoid
> that?
> Thanks a lot
>
|||I get those errors from time to time when I try to repliate over GPRS.
Normally it works fine. They happen at the begining of the synchronization.
At server side there's no message about them, the merge agent doesn't start.
I use SL Server CE 2.0 with SQL Server 2000 and SP3. The synchronization
works fine, and the URL od the agent os correct.
Thanks Viajy
"Vijay [MSFT]" <vijayts@.online.microsoft.com> escribi en el mensaje
news:e260Zfn1EHA.3596@.TK2MSFTNGP12.phx.gbl...
> Where and how do see this errors? Can you give more context?
> --
> This posting is provided "AS IS" with no warranties, and confers no
> rights.
> OR if you wish to include a script sample in your post please add "Use of
> included script samples are subject to the terms specified at
> http://www.microsoft.com/info/cpyright.htm"
> "Juan" <juan@.mailinator.com> wrote in message
> news:%23aF6Shg1EHA.2112@.TK2MSFTNGP15.phx.gbl...
> It
> avoid
>

got seconds but need minutes

I have a result that comes out in number of seconds, but need to see it converted to minutes and hours and seconds. Is there a convert function that would do this?
Thanks,
DanDECLARE @.x int

SELECT @.x = 3600 + 1800

SELECT @.x/60/60 AS Hours,CONVERT(Int,(@.x/60.00/60.00-@.x/60/60)*60) AS Seconds|||I'm rather fond of:SELECT Convert(CHAR(8), DateAdd(second, 45296, '0:00'), 8)-PatP|||I like it!

Simple, Elegant...sql

global dataset for n number of forms

hi friends ,

Is there anyway to share a single dataset for different forms.. coz i am repeatedly using the same query for different forms...

No, you cannot share a dataset among reports. Anyway, it doesnt make a big difference because the same query is going to run against the database even it is same for different reports.

If you are using SQL query text in your reports, consider creating a stored procedure for the query and use the stored procedure in all your reports.

Shyam

|||thank you shyam.......|||

Can you please mark the post as answer?