We want to generate new dashboards on-the-fly in our Cognos server, one for each of our customers. We want all dashboards to point to the same relational database, where each table has a column identifying the customer associated with the data, and use a data-tray filter in each dashboard to show only the data for the respective customer (we will only allow customers to access their respective dashboard with a read-only role, so that they cannot edit their dashboard) .
We were able to use the Cognos SDK to locate the "Exploration" object for the original dashboard and then create a copy of it, then modify the "specification" attribute in the new Exploration object to filter on the correct customer identifier, but somehow the cloned dashboard keeps showing the same data as the original dashboard from where it was copied.
Upon closer examination, we found out that if the dashboard owner (me) puts the new dashboard in edit mode, applies a random filter to any field in the data tray, then undoes the change, then the dashboard shows the correct data, matching the filter for the customer. Once the dashboard is saved, it works correctly from that point on.
Using the Cognos SDK we could see that the only difference between the specification attribute before and after that 'fix' is a 'shapingId" attribute, which somehow seems to be a clue for Cognos about which filtering criteria to use.
This may be an unusual scenario, but is any plan to support creation and modification of dashboards and/or data modules via Cognos SDK ?
------------------------------
Denilson Nastacio
------------------------------
#CognosAnalyticswithWatson