2 packages, cube?

[Jeka]

Hello guys, please help!

1.Is it possible to build one report out of the 2 different packages?

2. or build a report out of a cube and a package?

3. why would you build a cube (in transformer) and post for to cognos connect instead of building a package in a FM, can the cube work better?

thank-you for your input!

[josepherwin]

1.Is it possible to build one report out of the 2 different packages? No

2. or build a report out of a cube and a package? Yes
You will need to republish the package that you want to include the cube as well. Read my answer on section 3 for more info

3. why would you build a cube (in transformer) and post for to cognos connect instead of building a package in a FM, can the cube work better?
If you build cube and post to Cognos Connect, it will only create a package that consist of only the Cube itself.

If you publish it through a package in a FM, you will have the option of :
1. publishing multiple cubes under 1 package

2. publish a cube and other query subjects under 1 package

[Jeka]

Joseph,
Thank you for your reply.

last question:  :) What is the reason of building the cube, (i was told if you building dimension hierarchy it might Be better performance) cant you just build a package (and use the same hierarchy) with the same data source which the cube is using? or the cube can be better option in certain environment?

thank-you!

[josepherwin]

Cubes are normally used to perform analysis over a long time (let say 7 - 10 years of data)

By using cubes, you wont have to hit the database at all, it will directly hit the powercube.

If we are talking about a large amount of data, using powercube will definitely more efficient.

Imagine if 100 of your user is hitting the database trying to run a report based on 10 years worth of data on the lowest level of granularity and joining it with 5 or 6 other dimensions.

[Jeka]

Joseph, again thanks for your replies!

If i build cube with 5 levels of hierarchy and post  into cognos connect, use it in  report studio.
But than i have users wanting to use an Analysis Studio but i don't want him to see the certain accounts and other users i don't want him to see other accounts. departments only should see their own accounts!

Is it possible to set up the cube in the way i define what the user will see when he uses that cube? or i have to create cubes for every department?

Thank-you!

[josepherwin]

It's possible to do it. This functionality exist in Transformer 8.3 and it is called Custom Views.

As the name implies, it provides custom views of the cube based on your user group.

New in Transformer 8.3, you can even now prohibit certain users/groups from seeing certain measures.

[Ty Clabbers]

It is also possible in older versions of Framework. I don't know for sure if you need to authenticate with access manager but ithink so. Anyway if you use that as your ldap you can.
First you have to decide if you need authorisation within dimension data or just by dimension. If you need the first (ie every user should only see their own clients) you should first populate your dimension and then create views on that data (this can be done in the diagram viewer). You then go to the model properties (under the file menu), select the authentication tab and check include access manager user classes in the model. You now can go to the diagram viewer and select the user class tab. There you can use the tree structure to omit data for the several levels. Of course if you need to omit complete dimensions you can get away without creating the views. Then you can create your cubes the normal way.
If you are using connect connection you should then open the properties of the connection of the data source and select the authentocation to an external namespace.