1
votes

I am working on a project for work and both me and my boss are stumped. We have created a few SSAS cubes to work with views in our database. 6 in total. All seprate. The first 5 work just fine with all cubes being set up the same. The last one we set up has an issue. The measures are showing up wrong. Not the grand total measures. I know that is a problem for some people. But the individual measures.

For example we have a view that has the unique key 'Record ID' which is unique in the table. However in the cube the count for some, but not all ID's are showing up as 2 or more. Along with that another measure we use 'Work Hours' shows up incorrect as well. In our database, as well as if I 'Explore Data' on the table in the DSV, a particular record ID may show up with 2 work hours. But in the cube it shows up with 12.

This cube being set up the same way as the other cubes, which is partly why im confused because the other ones work just fine.

Things I have thought:

The cube is summing/averaging items that have the same information other than Record ID. This cant be it because adding up or averaging the hours for similar items doesnt provide the work hours given.

The cube is getting incorrect data. Cannot be this either because the data in the data source view is correct. The measure does not get run through the dimension, when I do add the measure in the dimension that value shows up correct. (However we need to use these numbers as a measure not as part of the dimension)

Additional information; From what I've seen common issues are that the Calculate is missing, in this case it is not.

The measures are set up incorrecly. I doubt this is the case because I am pulling them into the wizard the same way as our previous cubes but not getting proper results.

Any information would prove useful to us. We have been going through this process for a week now. Thanks again in advance!

2

2 Answers

1
votes

Actually I figured it out. Not sure why it was doing this but what the cube and dimension were doing was they were taking certain records and deleting them, taking the measures from them and adding it to the record before it. I know that seems like its not the case but when I went to browse a certain record, say one ending 1996, (not a year attribute), its work hours would be 2 in the database. But in the cube work hours would be 2.2. That .2 being the work hours from the next record ending 1997, which is not in the cube but is in the database.

To fix this the column was changed from a double to BigInt considering the ID's never had a decimal.

0
votes

Have you included any Dimensions in the new cube that are not in the other 5?

When I've had multiplicity in cubes it's usually boiled down to duplicate keys. I'd suggest focus on one specific record ID in the database and join to each dimension in turn.hopefully this will indicate where the problem lies.