Aggregation - Incorrect results

All,

I have noticed a behaviour in SSAS 2005 that I would like to put forward to understand it better and to know if anyone else have encountered similar issues.

I have Cube with 4 measure groups (with only 1 visible to user. Others are for implentation for Many to Many dimension and Currency conversion). The measure group visible to user has 1 partitions initially and other are created on fly using AMO. The aggregation design is cloned from the initial partition. When I create aggregation design using the inbuilt aggregation design wizard, it is working fine. But when I use the Agg manager utility from the BIDS Helper addin and create agg from querylog and process the cube, the result returned from the cube is incorrect.

My question : Why does an aggregation making the cube to return incorrect result?

Cheers,

Arun

Hi Arun,

Does your aggregation include an attribute from a non-materialized reference dimension (see issue below)?

...

To make a long story short, I had manually designed aggregations that I shouldn't have. We have a reference dimension with materialized set to False. This reference dimension should not have been included in any aggregations. The Aggregation Design Wizard is smart enough not to do this, however you can hack the XMLA to do so yourself. The Aggregation Manager Utility included in SP2 also allows one to add aggregations to these type of dimensions.

...

What the correct behavior should be

When Materialize is set to False for reference dimension, aggregation should not include any attribute from the reference dimension. The correct behavior should be the processing would raise an metadata validation error. A design change request to implement this has been filed to the next version of Analysis Services.

What is the problem

With this aggregation included in the cube, the aggregation creation routine does not know how to aggregate the measure from fact and the reference dimension attribute as the fact does not have the attribute from the reference dimension, therefore the measure is null for this aggregation.

...

Aggregate integrity problems

Has anyone else had any problems with cubes showing incorrect aggregates?
For instance, the sales for a Quarter is showing twice the amount of the 3
months showing. As well, there are values associated to products (lowest
level in dimension) that should not be there. When I re-process the cubes,
the issues are eventually resolved but I feel that I cannot trust any numbers
that the cubes show now.
Thanks in advance
Data Integrity starts at the souce and, once validated, is only maintained
through constraints.
The old adage applies here: Garbage In, Garbage Out.
I'd look at the integrity constraints of the primary sources. If those look
intact, I'd attempt to run manual queries to aggregate the data to see if any
anomolies show up. Finally, if all looks well, the next time some
abnormality shows up in the cube, instead of reprocessing, I'd look at the
source first to see if you could manually replicate the numbers.
I have not seen what you are describing except where there were data errors
in the sources. But if you feel that all has been checked out, I would most
certainly open up a PSS case.
Sincerely,
Anthony Thomas
"fastforward" wrote:

> Has anyone else had any problems with cubes showing incorrect aggregates?
> For instance, the sales for a Quarter is showing twice the amount of the 3
> months showing. As well, there are values associated to products (lowest
> level in dimension) that should not be there. When I re-process the cubes,
> the issues are eventually resolved but I feel that I cannot trust any numbers
> that the cubes show now.
> Thanks in advance
|||I agree, Garbage In, Garbage Out. I have experience more than I want to.
In this case however, the data is clean. It has been checked and double
checked. There are no problems with this set of data. Could this have
anything to do with how I have designed my aggregates or writeback issues?
"AnthonyThomas" wrote:
[vbcol=seagreen]
> Data Integrity starts at the souce and, once validated, is only maintained
> through constraints.
> The old adage applies here: Garbage In, Garbage Out.
> I'd look at the integrity constraints of the primary sources. If those look
> intact, I'd attempt to run manual queries to aggregate the data to see if any
> anomolies show up. Finally, if all looks well, the next time some
> abnormality shows up in the cube, instead of reprocessing, I'd look at the
> source first to see if you could manually replicate the numbers.
> I have not seen what you are describing except where there were data errors
> in the sources. But if you feel that all has been checked out, I would most
> certainly open up a PSS case.
> Sincerely,
>
> Anthony Thomas
>
> "fastforward" wrote:
|||Data integrity is always issue number one but if you've checked that out,
then logic comes next. I'd run the aggregate process manually on a few
deminsions with normal T-SQL aggregates. Then, most certainly, I would
consider a more simple design for your cubes to see if the problem clears up.
If so, I'd add complexity, one layer at a time.
If you continue to run into aggregations issues, I would most certainly
consider contacting MS PSS.
Sincerely,
Anthony Thomas
"fastforward" wrote:
[vbcol=seagreen]
> I agree, Garbage In, Garbage Out. I have experience more than I want to.
> In this case however, the data is clean. It has been checked and double
> checked. There are no problems with this set of data. Could this have
> anything to do with how I have designed my aggregates or writeback issues?
> "AnthonyThomas" wrote:

Aggregate integrity problems

Has anyone else had any problems with cubes showing incorrect aggregates?
For instance, the sales for a Quarter is showing twice the amount of the 3
months showing. As well, there are values associated to products (lowest
level in dimension) that should not be there. When I re-process the cubes,
the issues are eventually resolved but I feel that I cannot trust any numbers
that the cubes show now.
Thanks in advanceData Integrity starts at the souce and, once validated, is only maintained
through constraints.
The old adage applies here: Garbage In, Garbage Out.
I'd look at the integrity constraints of the primary sources. If those look
intact, I'd attempt to run manual queries to aggregate the data to see if any
anomolies show up. Finally, if all looks well, the next time some
abnormality shows up in the cube, instead of reprocessing, I'd look at the
source first to see if you could manually replicate the numbers.
I have not seen what you are describing except where there were data errors
in the sources. But if you feel that all has been checked out, I would most
certainly open up a PSS case.
Sincerely,
Anthony Thomas
"fastforward" wrote:
> Has anyone else had any problems with cubes showing incorrect aggregates?
> For instance, the sales for a Quarter is showing twice the amount of the 3
> months showing. As well, there are values associated to products (lowest
> level in dimension) that should not be there. When I re-process the cubes,
> the issues are eventually resolved but I feel that I cannot trust any numbers
> that the cubes show now.
> Thanks in advance|||I agree, Garbage In, Garbage Out. I have experience more than I want to.
In this case however, the data is clean. It has been checked and double
checked. There are no problems with this set of data. Could this have
anything to do with how I have designed my aggregates or writeback issues?
"AnthonyThomas" wrote:
> Data Integrity starts at the souce and, once validated, is only maintained
> through constraints.
> The old adage applies here: Garbage In, Garbage Out.
> I'd look at the integrity constraints of the primary sources. If those look
> intact, I'd attempt to run manual queries to aggregate the data to see if any
> anomolies show up. Finally, if all looks well, the next time some
> abnormality shows up in the cube, instead of reprocessing, I'd look at the
> source first to see if you could manually replicate the numbers.
> I have not seen what you are describing except where there were data errors
> in the sources. But if you feel that all has been checked out, I would most
> certainly open up a PSS case.
> Sincerely,
>
> Anthony Thomas
>
> "fastforward" wrote:
> > Has anyone else had any problems with cubes showing incorrect aggregates?
> > For instance, the sales for a Quarter is showing twice the amount of the 3
> > months showing. As well, there are values associated to products (lowest
> > level in dimension) that should not be there. When I re-process the cubes,
> > the issues are eventually resolved but I feel that I cannot trust any numbers
> > that the cubes show now.
> >
> > Thanks in advance|||Data integrity is always issue number one but if you've checked that out,
then logic comes next. I'd run the aggregate process manually on a few
deminsions with normal T-SQL aggregates. Then, most certainly, I would
consider a more simple design for your cubes to see if the problem clears up.
If so, I'd add complexity, one layer at a time.
If you continue to run into aggregations issues, I would most certainly
consider contacting MS PSS.
Sincerely,
Anthony Thomas
"fastforward" wrote:
> I agree, Garbage In, Garbage Out. I have experience more than I want to.
> In this case however, the data is clean. It has been checked and double
> checked. There are no problems with this set of data. Could this have
> anything to do with how I have designed my aggregates or writeback issues?
> "AnthonyThomas" wrote:
> > Data Integrity starts at the souce and, once validated, is only maintained
> > through constraints.
> >
> > The old adage applies here: Garbage In, Garbage Out.
> >
> > I'd look at the integrity constraints of the primary sources. If those look
> > intact, I'd attempt to run manual queries to aggregate the data to see if any
> > anomolies show up. Finally, if all looks well, the next time some
> > abnormality shows up in the cube, instead of reprocessing, I'd look at the
> > source first to see if you could manually replicate the numbers.
> >
> > I have not seen what you are describing except where there were data errors
> > in the sources. But if you feel that all has been checked out, I would most
> > certainly open up a PSS case.
> >
> > Sincerely,
> >
> >
> > Anthony Thomas
> >
> >
> > "fastforward" wrote:
> >
> > > Has anyone else had any problems with cubes showing incorrect aggregates?
> > > For instance, the sales for a Quarter is showing twice the amount of the 3
> > > months showing. As well, there are values associated to products (lowest
> > > level in dimension) that should not be there. When I re-process the cubes,
> > > the issues are eventually resolved but I feel that I cannot trust any numbers
> > > that the cubes show now.
> > >
> > > Thanks in advance

Aggregate integrity problems

Has anyone else had any problems with cubes showing incorrect aggregates?
For instance, the sales for a Quarter is showing twice the amount of the 3
months showing. As well, there are values associated to products (lowest
level in dimension) that should not be there. When I re-process the cubes,
the issues are eventually resolved but I feel that I cannot trust any number
s
that the cubes show now.
Thanks in advanceData Integrity starts at the souce and, once validated, is only maintained
through constraints.
The old adage applies here: Garbage In, Garbage Out.
I'd look at the integrity constraints of the primary sources. If those look
intact, I'd attempt to run manual queries to aggregate the data to see if an
y
anomolies show up. Finally, if all looks well, the next time some
abnormality shows up in the cube, instead of reprocessing, I'd look at the
source first to see if you could manually replicate the numbers.
I have not seen what you are describing except where there were data errors
in the sources. But if you feel that all has been checked out, I would most
certainly open up a PSS case.
Sincerely,
Anthony Thomas
"fastforward" wrote:

> Has anyone else had any problems with cubes showing incorrect aggregates?
> For instance, the sales for a Quarter is showing twice the amount of the 3
> months showing. As well, there are values associated to products (lowest
> level in dimension) that should not be there. When I re-process the cubes
,
> the issues are eventually resolved but I feel that I cannot trust any numb
ers
> that the cubes show now.
> Thanks in advance|||I agree, Garbage In, Garbage Out. I have experience more than I want to.
In this case however, the data is clean. It has been checked and double
checked. There are no problems with this set of data. Could this have
anything to do with how I have designed my aggregates or writeback issues?
"AnthonyThomas" wrote:
[vbcol=seagreen]
> Data Integrity starts at the souce and, once validated, is only maintained
> through constraints.
> The old adage applies here: Garbage In, Garbage Out.
> I'd look at the integrity constraints of the primary sources. If those lo
ok
> intact, I'd attempt to run manual queries to aggregate the data to see if
any
> anomolies show up. Finally, if all looks well, the next time some
> abnormality shows up in the cube, instead of reprocessing, I'd look at the
> source first to see if you could manually replicate the numbers.
> I have not seen what you are describing except where there were data error
s
> in the sources. But if you feel that all has been checked out, I would mo
st
> certainly open up a PSS case.
> Sincerely,
>
> Anthony Thomas
>
> "fastforward" wrote:
>|||Data integrity is always issue number one but if you've checked that out,
then logic comes next. I'd run the aggregate process manually on a few
deminsions with normal T-SQL aggregates. Then, most certainly, I would
consider a more simple design for your cubes to see if the problem clears up
.
If so, I'd add complexity, one layer at a time.
If you continue to run into aggregations issues, I would most certainly
consider contacting MS PSS.
Sincerely,
Anthony Thomas
"fastforward" wrote:
[vbcol=seagreen]
> I agree, Garbage In, Garbage Out. I have experience more than I want to.
> In this case however, the data is clean. It has been checked and double
> checked. There are no problems with this set of data. Could this have
> anything to do with how I have designed my aggregates or writeback issues?
> "AnthonyThomas" wrote:
>