Feed Data into multiple elements through TI

[vinnusea]

Hi
AM trying to load data into a cube and Now i have a Issue.
Have a Dimension with usernames ( Elements) and am able to load all Units into username 'alpha' .
Now 'Alpha' is populated with all units for all periods..

Task:

I have other users 'Beta','Gama',..... I want to load same Units into all Others users using TI process. (Why not Rules is , i dont want UNITS for all users to become uneditable ) I have created a new varible name 'Vusers' and in formula wrote 'Vusers' = 'Alpha'; this formula is diverting my units to 'alpha' user. But now i need to write all other users... How is that possible???????

Main AIM: Alll Users should be able to Edit Units

Thank you for your time!!!!

[Martin Ryan]

Let me see if I understand you correctly, you are trying to load the same data against users so that all the users can edit the data? So do you mean you have the }Clients dimension as part of your cube?

Unless you have a really weird requirement this is not the way to go about security. You should not have the }Clients dimension in any cube.

You should load data into a cube structure that reflects the data, completely ignoring users and security. Once you have the data loaded then you start thinking about security, and you do this by giving users read/write/none access to a key dimension. Often this is something like the Cost Centre dimension. E.g. Alpha will have access to HR, Beta will have access to IT and Gamma will have access to Finance.

To learn about security refer to the manuals, but the very brief version is right click on the dimension (or dimensions) you want to use for security, go to security and choose Elements Security Assignments...

Martin

[vinnusea]

Yes am using }Groups Dimension in my Cube. So is that possible to load data to all Groups so that each person in each group can edit it...

[tomok]

Yes am using }Groups Dimension in my Cube. So is that possible to load data to all Groups so that each person in each group can edit it...

Regardless of whether or not this is a good idea you can certainly write the same data to each element in the }Groups cube as long as the }Groups dimension is part of the cube. You can either write a separate CellPutN statement for each group in your Data tab or alternatively do a WHILE/END thingy in the Data tab, looping through the }Groups dimensions and writing the data to each group.

[vinnusea]

okay ..
Am using a sql for pulling data to load into cube which was having 6 columns in sql . I dont have any Groups coming from my source. So i defined a New Varible.
For writing a While loop , here in Variables Tab in TI what do i need to define my Groups Varible ..?

What formula i need to use in varible tab for my groups varible????


Thanks for reply.. the while loop will work i think..

[Martin Ryan]

Are you absolutely certain you want to have }Groups in your data cube? It is really not a good idea. I can vaguely conceive where it might be useful if you wanted to track all the changes multiple users were making, but in that instance I would use a Scenario dimension, not the Groups dimension. I would never ever use the Groups dimension in data cube.

If you are really, absolutely, 100% certain that you want }Groups in your data cube then then TI syntax is something like

Code: Select all

numGroups=dimsiz('}Groups');
i=0;
while(i<numGroups);
i=i+1;
group=dimnm('}Groups', i);
cellputn(value, 'dataCube', dim1, dim2, dim3);
end;

Martin

[vinnusea]

Thanks Martin.
So i really dont need a }Groups Dim But why i took it .. Our users need to do their own What if analysis so am using 9.4 so dont have sand boxes. So i thought to create a dimension with user names who ever logs in they ll see thier own data.

Can you help me if i need a Dimension which will dynamicly update when their is a new Group or Client added in server.. so that i can remove this }Groups and use the Dimension .
If any TI process can help me for updating the Dim with New Groups that adds in Server ...........

Really Appriciate your help

[Martin Ryan]

The idea behind sandboxes in 9.5.1 is to have an area to try things out before submitting them for general consumption. It's not supposed to be a permanent record of various what-if analysis. (Its other purpose is to cope with Cognos Planning users who are used to "submitting" data, rather than the usual way of numbers going back to the server as soon as they're entered).

Can you please describe overall what you're trying to achieve? What data are you importing from SQL? Usually data imported from ODBC sources is actual data, but if you want to apply what-if scenarios to it, that suggests it contains forecast data, is that right?

What kind of data are you analysing here? Sales numbers? Financial? Metrics?

Are you users supposed to work together, or are they doing their own what-if analysis completely independent of one another? Do multiple users/groups have access to the same key area (e.g. cost centre)? If so, why?

As a general rule for what if analysis I would have a handful of "What If N" scenarios (e.g. "What If 1", "What If 2", "What If 3", I might have aliases to indicate what they are, e.g. "Takeover by Facebook", "Recession", "Massive growth"). Because users should be looking after their own products or cost centres then several users can use the same what if scenario - so there is no need to create a scenario for each person. Creating a lot of copies of the data will blow out the cube size which is likely to slow things down (c.f. http://forums.olapforums.com/viewtopic.php?f=3&t=3661).

Martin

[paulsimon]

Vinnuesa

I use a routine that copies the }Groups dimension to a zTD_Control_Groups dimension. I then use that in cubes. The advantage is that you can then create consolidations etc. You can also remove groups from the system which changes the }Groups dimension without necessarily losing their details from the zTD_Control_Groups dimension.

However, this dimension won't automatically update when you add a new Group. But then even if you continue to use }Groups you will need to run a TI process to populate your initial value.

I suggest that you need to write a TI process to add a Group. In our system we always create a personal group for each user. We use a TI process that accepts the Client Name. It automatically creates the Clients, generates a Group name by adding 'Grp_' to the Client Name, and puts the Client in to that Group, along with other standard Groups that give a basic level of access to the system. Other access is then controlled by other procedures. Part of the procedures run the routine that updates the zTD_Control_Clients and zTD_Control_Groups dimension.

If you cannot do all this at the time that you create the Client and Group, then you could still copy your }Groups dimension to the zTD_Control_Groups dimension just before you update the cube with the values that you read in from your relational table.

Regards


Paul Simon

[vinnusea]

Yes
Users doing what if analysis on forecast data. All users get same data from data base which is having forecasted data uploaded from forecast cube in other server. In new server they have new measures .New measures Input will be independent to each user so each user Input his own data and checks the Final values and do what if.
That is the reason i want to load data from DB to each user so that they gets the data and edits it to their interest.
-- Thanks