What Is Framework Manager?

Framework Manager stores all sorts of information about your data in a Framework model. This process is known as metadata modelling. Using Framework Manager items from several database tables can be collected into a single item (known as a query subject), given user-friendly and meaningful names and then ‘published’ to Cognos Connection from where users and report developers can use it to produce reports and analyses. A test function allows the modeller to see how the data will look when it is published.
Once information such as joins and calculations is published it can be used again and again, with no risk of getting the joins wrong. Potential reporting errors known as ‘reporting traps’ can be resolved in the Framework model to avoid these errors being made in reports or other data analysis.

Framework Manager has a sophisticated security system, so you can restrict sensitive items to particular users, groups of users or users assigned to particular ‘roles’, for example Administrators.

Data can be modelled relationally, similar to how database tables are joined together (under the surface of course), or dimensionally, which means data can be represented in hierarchies with levels. For example, modelling a date dimension allows the user to view summarised data at Year level, then drill down see it summarised at Quarter, Month and even Day level if required.

Calculations can be made between different data items using a wide range of functions and then stored in the Framework model so everyone can access them. Filters can be built into the model, and prompts created in Framework Manager are available to any number of reports and analyses through the published package.

The ability to rename data items in terms that are meaningful to the data users is an important feature of Framework Manager. You can also define the type of data (Framework Manager will try to guess if you don’t), and this will affect how the data item is processed in a report or analysis. For example if an item is defined as a ‘fact’ it will automatically be summed.

Once modelled data can be packaged in several ways depending on its function.

A well-written and tested Framework model makes reporting faster and avoids potential errors and duplication, as the same package can be used in any number of reports and analyses.