August 7, 2014 at 2:31 am
Dear Colleagues,
Since some days ago I could not find data in some measure groups in my cube in the dev
system, even when I am able to see them using the "explore data"
option in the cube structure.
This behavior is only in the dev server, fortunately the cube in the test server works fine.
I deploy always the same version of the cube using TFS.
Obviously is not the "CALCULATE", at the beginning of the Calculations, and
everything point to the source data in the dev server.
The data in the test server is closed to the real one and it is loaded everyday while the
data in the dev server is test data and is not reloaded since some days ago.
Have anybody experienced a similar situation?
Any comments would be extremely appreciated.
Kind Regards,
August 7, 2014 at 2:33 am
Any filters in the data source view?
Need an answer? No, you need a question
My blog at https://sqlkover.com.
MCSE Business Intelligence - Microsoft Data Platform MVP
August 7, 2014 at 2:59 am
Hi Koen,
thanks for your quick answer.
I am not applying any filters in the dsv (actually I don’t know how to do that), in fact I manipulate the data only in the view layer and in the dsv I just bring the table and define the relationships.
What I cannot understand is why the same cube works ok in the test server.
I maintain the cubes and the database structure using TFS.
Kind Regards,
August 7, 2014 at 3:03 am
In Visual Studio which version of the data are your project data sources looking at?
Are you using solution configurations?
Where are you exploring data, in the dsv or in the dimension view?
Does the cube process? When it processes is the data there as expected?
Sorry for all of the questions! I've found over the years that Visual Studio (or BIDS or SSDT, whatever!) can behave unusually in this regard, particularly if you are using an "offline" project rather than connecting directly to a processed AS database. When you are using different data sources it can get confusing if you are using solution configurations (the cube itself may be the cube on test whereas the dsv might be looking at test for example). It can also get confusing if you edit a cube using a direct connection and then go back to the edit>deploy>process model. Some stuff can be edited directly without the need to process (MDX calculations for example).
Verify what you have deployed and processed is correct by looking at the data sources in SSMS. If they are incorrect then change them. Get it all "in line" 😀
August 7, 2014 at 3:41 am
Hi,
please find below my comments:
yayomayn (8/7/2014)
In Visual Studio which version of the data are your project data sources looking at?I have 2 environments, one dev server and one test server, both with the same version of sql server and analysis services. The difference is only the content of the databases. In the dev server I have sample data and the quantity is also smaller als in the test server.
Are you using solution configurations?
The solution configuration is the default one: "Development". What I do is check the last changes in TFS and the open visual studio with another user that has rights in the test server, download the last version and deploy the cube.
Where are you exploring data, in the dsv or in the dimension view?
I am exploring the data in the Cube Structure Tab of the .cube file, but if I do it in the dsv I also get data. If I browse the cube in the Browser Tab there are no data and I get the message "No rows found. Click to execute the query"Does the cube process? When it processes is the data there as expected?
The cube process without error, but in 10 seconds, even if I force a full process.
Sorry for all of the questions! I've found over the years that Visual Studio (or BIDS or SSDT, whatever!) can behave unusually in this regard, particularly if you are using an "offline" project rather than connecting directly to a processed AS database. When you are using different data sources it can get confusing if you are using solution configurations (the cube itself may be the cube on test whereas the dsv might be looking at test for example). It can also get confusing if you edit a cube using a direct connection and then go back to the edit>deploy>process model. Some stuff can be edited directly without the need to process (MDX calculations for example).
We have a separated user for every environment, and we use windows authentication, therefore if I open visual studio with my dev user I cannot access the test server and viceversa.
Thanks for your detailed answer, I think I doing something with the change of users and TFS.
Kind Regards,
Verify what you have deployed and processed is correct by looking at the data sources in SSMS. If they are incorrect then change them. Get it all "in line" 😀
August 7, 2014 at 4:18 am
I think that the issue is that you only have one version of the project and since you are using different users to open these projects one might have permissions where the other doesn't. I'm no expert on TFS, but I think you can get around this by "branching" the project so you have a test version and a development version. These can in turn have their own data source connections etc.
To verify that the cubes are there and correct, browse them directly in SSMS. If they are fine then the above is your problem.
Regards
Paul 🙂
August 18, 2014 at 3:06 am
Hi Friends,
it was really really strange, but I had to reload the data mart database and then my data is back.
It seems something like SSAS was not able to recognized that the measure groups should be processed, even if I select the option "Process Full" and a kind of incremental or delta process was performed instead, and because there were no new data, nothing was processed.
I know my explanation is a little bit confused but like all this issue.
I can only say that reload the database solved my problem.
Kind Regards,
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply