Another Time Dimension question

I want to track sales vs. both date scheduled to be shipped and date actually shipped. Both are fields in my OLTP. I can set the SSIS program to extract data any why I choose.

Would it be best to create a fact table and 2 different dimension tables or combine the time fields in one dimension table or leave them in the fact table and let SQL Server extract the time dimension?

Thanks.

Hello! In SSAS2005(Analysis services 2005) you have something called 'Role-playing' dimensions that solves this problem. You will only need one time dimension in the starschema that you join to all dates in the fact table.

If you have these relations designed in the source system/star schema or in the data source view SSAS2005 will detect this relation and create separate time dimensions automatically when you build the cube.

HTH

Thomas Ivarsson

|||I don't want to apply one time dimension to many fact tables. I want to apply many time dimensions to one fact table. Right now my date fields are in the Fact Table.|||

I do not think that I have said that but my explanation, perhaps, was not good enough.

You use one time dimension and join the fact tables different date dimensions to the same date key in the time dimension table.

Connect the key for order date in the fact table to the date key in the time dimension. Connect the date key for invoice date in the fact table to the date key in the time dimension.

HTH

Thomas Ivarsson

|||

Ahh...that's better....I think I need to take my dates out of my fact table and create a dimTime table and then I can create the multiple relationships.

Thanks.

|||

That seems to be working but I'm a little fuzzy on the theory.

Either or both date fields may be null in my OLTP table, so I can't use either for the key field. My OLTP table has an integer key that is just a counting number.

I created the dimTime table with the same type key field. When I did my SSIS run I just copied the Fact table key field to the dimTime table key field along with the date fields.

The tables are linked but not throught a date field. SSAS is smart enough to build its aggregates on the date fields and pretty much ignore the key field? It only cares that the tables are linked somehow?

Thanks!!!!

|||

You can add a theoretical time member to the time dimension like '2099-12-31" and point fact records without time members to that time member.

There will be many more SQL Server releases until we reach that date.

HTH

/Thomas Ivarsson

Analysis Services Roles - Dimension Data

Is there a way to assign dimension data member restrictions within AS 2005 Roles to all attribute hierarchies within a dimension

For example we have a dimension with 4 attribute hierarchies :

Office Code

Office Name

Office Group

Office Overseas

Office Code and Office Name roll up to Office Group (within Office Overseas)

We want to restrict certain Active Directory groups to only see their own office data eg Office Code TOR for Office Name Toronto (within Office Group Canada)

When an AD group is assigned to only see Office Code TOR (by using the Role Dimension Data member restrictions)

Then users in this AD group can still see measure data aggregated for all Office Names and Office Groups (though can only see Office Code TOR)

Do restrictions have to applied separately to all attribute hierarchies within a dimension? This will be difficult for us to maintain, since most of our dimensions are type 1 with members moving (eg the office groups change quite frequently)

Has anyone else come across this issue ?

>Then users in this AD group can still see measure data

>aggregated for all Office Names and Office Groups

>(though can only see Office Code TOR)

But the realy see Values for Toronto, not for all the offices, don't they?

|||

I'm not sure I'm 100% clear on the problem you are having. Are you saying the dimension members are appropriately restricted but the totals reflect non-visible members? For example, if you drill down to Office Code you see just TOR. The total for All Office Code is different than the TOR value?

If this is what you are trying to address, take a look at the Visual Totals setting in your dimension data security. By setting this value to True, SSAS calculate the totals based on just those members you have access to.

Thanks,
Bryan Smith

|||

thanks Bryan

setting the visual totals on the dimension data assignment has resolved this

Mark

Analysis Services on SQL2000 SP4

"unable to browse the dimension [dimension name].

Provider cannot be found. or Unknown Error.

Everything else looks good, like processsing, looking at the tables,fields, datasource is connected.

Dear,

Did u used a shared dimension in the cube.

from

sufian

|||

I tried both, shared and not shared. Input is from Access db. It fails already in the Create Dimension Wizard screen. Data should be populated on the screen

|||

Can you please describe your situation in bit more details.

What version of Analysis Services you are using?
What relational database are you using?
What OLEDB provider ?
What is the exact error message you are getting and at what stage?

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

|||

Analysis Services Version 8.0.194

Input Database MS Access2000

Provider = MS Jet 4.0 OLE DB Provider

OS = Windows 2000 Professional with SP4

Error appears already on the Finish the Dimension Wizard Screen

"Unable to browse the dimension [dimension name]. Unspecified Error.

I have processed the cube and the dimension. I get the count on all records processed.

Also if I try to browse data on individual dimension(s)I get ;

"Unable to browse the dimension [dimension name]. Provider not found or not properly installed" however testing the database connection works fine.

|||

What happens if you try to browse your cube after your processed it? Do you see dimensions and measures?

I suspect you just having problem browsing dimension. Is that so?

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

|||

FYI - I am using Professional SQL Server 2000 Programming by Robert Vieira (WROX) See page 920 top(data is already displayable for the Customer Dimension at creation time. I fail right there.

To answer your questions: I cannot browse the cube after processing - Same error "Unable to browse the cube - unspecified error".

Same applies to the dimensions.

Validate dimension, Count Dimension members and metadata all check out. Processing is going to normal end for both cube and individual dimensions and I get the correct number of members that I am supposed to get with no errors in processing.

|||

It is possible that you got some client components installation corrupted.

Try installing "Microsoft SQL Server 2000 PivotTable Services " from http://www.microsoft.com/downloads/details.aspx?familyid=d09c1d60-a13c-4479-9b91-9e8b9d835cdc&displaylang=en

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

|||

Thanks again, but no cigar. To install pivot tables I had to de-install Analysis Services, Install the pivot tables and re-install the Analysis. However I can't browse the results and/or the cube.

I went a step further installing SQL Server 2000 and Analysis Services on a different PC, using Windows XP with the same negative results.

Unable to browse the Cube [cube name] or

Unable to browse the dimension [dimension name]

Unspecified error. It will help how to find and read the OLAP server error logs or dump for a better understanding of what is wrong.

Since I am using the same version of the SQL Server (developer edition 8.00.760(SP3) from Microosoft) on both PC's I tend to believe there is a limitation on it. I have no other logical explanation.

|||

Are you talking about SP3 or SP4? The title suggests you'd like to install SP4.

Is that so?

Please make sure your run the same service pack version of the server and client.

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

|||

Veo,

Is the Access database stored on a network drive? The Analysis Services service would normally be installed to run as a system account which would not have access to network resources.

|||

I am the administrator of my PC. Acces is installed on the same PC with the SQL Server.

Operating system Windows2000 has SP4 applied and SQL Server has SP3 applied.

My next step will be to import Access data base into SQL 2000 to see if that makes a difference

|||

I think you should be better off with keeping your database in SQL Server.

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

Analysis Services help - Dim builds, but cube fails.

If there are any AS gurus out there, I could use some help. I've been having some problems with particular AS dimension and cubes and it's driving me crazy! It doesn't matter what I do, nothing seems to work.

Anyway, here's what I'm trying to do. I've got a fairly simple dimension. There is a date stored in the dimension that is formatted as an int. The dimension needs to only display AGE, so I cast the int as a date and do a datediff to get AGE. The dimension builds just fine and I get the results I want. My problem is when I add the dimension to the cube. The cube fails to build and I get an error message - "Data source provider error: The column prefix 'MY TABLE' does not match with a table name or alias used in the query." Basically, AS is telling me that the table that is used for the dimension doesn't exist, even though the fact and the dimension are joined properly and I've validated the structure.

I've run a number of queries in QA on the two tables and everything works fine. No funky data issues. I've run the service packs a few times, but that didn't work either. I've tried making a cube that only has the just fact table and the one dimension table, and it still fails.

Basically, I'm out of ideas. Any help that anyone has is greatly appreciated.My first suggestion would be to examine Brett's excellent sticky (http://www.dbforums.com/t1196943.html) with suggestions for getting quick answers. You certainly appear to have done your homework in understanding the problem, but haven't done much to help me understand it!

If that error message is literally what AS reports, I suspect that the problem is the space in the 'MY TABLE' name. If you use characters outside of letters, digits, and underscores in your object names (like your space), then you need to use square brackets [] to surround that name.

-PatP|||My first suggestion would be to examine Brett's excellent sticky (http://www.dbforums.com/t1196943.html) with suggestions for getting quick answers. You certainly appear to have done your homework in understanding the problem, but haven't done much to help me understand it!

If that error message is literally what AS reports, I suspect that the problem is the space in the 'MY TABLE' name. If you use characters outside of letters, digits, and underscores in your object names (like your space), then you need to use square brackets [] to surround that name.

-PatP

Sorry. The table isn't actually called 'MY TABLE'. I was just using that as a generic reference. The real table name is 'F_LOAN'.

Pat, do you have any specifc questions I can answer? Are you pretty familiar with AS? In a nutshell, I have a dimesion that builds just fine. When I add it to the cube, the cube fails to build. I'm guessing that I'm completely missing something very basic or I've got a bug. :eek:

I've added some screen shots of the cube, the error and the dimension. A pic is worth a thousand words, right?|||A little more info...

If I remove the dimension logic that calcs the Loan Age, the cube will build. However, I don't understand how the logic could be a problem when the dim builds fine with the logic. I would expect the dimension to fail if there was a problem with the Loan Age calc logic.|||One more pic. This query is doing what I want the dimension to do.|||Well, I think I've got it solved. I know everyone was waiting with bated breath! :D

Evidently, the cube didn't like having a dim where the dim converted the date, stored as an int in the DB, to a date. Even though the dim was working just fine, the cube somehow couldn't handle it. So, I made a view based on my dim table and did the int-to-date conversion there. Then I used the view as the source for my dim. And...everything worked just fine. I guess I was just asking too much from AS.

Thanks.

Analysis Services Dimension Question

I have a database with millions of users with email addresses. I want to create an email domain dimension that groups domains into all the big email domains (hotmail, aol, yahoo) and an "other" consisting of all the other domains. I can create a table with entries for the big domains, and get the grouping working but anything that is not a big domain will get thrown out rather than put into an "other" category.

Any idea how to get such an "other" category?I just love SPAM. May I have another serving please ?

-PatP

Analysis Services Developer Studio Question

I had an obsolete dimension on my cube.

I delete the dimension, it disappears from the GUI (I can see all the other dimensions in the dimensions panel, but this one correctly disappears), yet when I try to build my project I get errors on this dimension. It's like the dimension was deleted from the GUI but still exists somewhere and is causing problems.

How do I resolve this without rebuilding my cube entirely? How do I delete this dimension completely?Are you sure that dimension is absolete ?
If it is not - error looks correct ...
What error do you receive ?

Analysis Services Data Source Cube vs Dim

Hi group

I am creating a cube where the facr table is held in one sql database/data source and the dimension info in another sql database/data source.

When I try to include the dimension in the cube the dimension is not available.

How do I get around this?

Thanks in advance

ChrisHave you tried creating virtual dimensions over your existing dimensions?