By Elwood Philbrick, Envisn, Inc.
Administering Cognos Analytics is mostly about managing change. Changes to reports, packages, models, security, triggers, schedules, etc. are ongoing. Doing this successfully requires being able to accurately understand how planned changes will impact the objects in your Content Store. Knowing the precise impact of planned changes is critical because if you don’t get this right the change process can be painful. Because objects requiring change are not identified or objects are incorrectly identified as needing change. Fixing them is difficult because you may not even know what’s been changed.
Identifying changes that will impact current objects in the Content Store is just the first step in the process of change management. Ideally we would have tools that not only do impact analysis, but also then make the necessary changes, and then validates that the changes were correctly made. Impact analysis itself is only the initial step in this process.
The most common areas where impact assessment is needed in Cognos include:
- Database dependency – All or part of a database is changing
- Model item dependency – Changes being made to an FM Model
- Package dependency – Impact of proposed changes to a package or content
- Data source dependency – Typically a single database but it can be more than one
- Specification dependency – Changes being made to objects beyond the data itself.
- Cube data dependency – PowerPlay Cube data sources being used by reporting objects
Model Item Dependency
A use case example of impact analysis that relates to model item dependency is where a change has been made to a model item labeled “Discount Amount” that will be replaced with “Special Discount Amount”.
We will use the Model Item change function to identify wherever the item Discount is used in any reports or queries and replace it with Special Discount Amount (Figure 1).
In Figure 2 we see the results of this impact analysis. We have identified 8 reports that contain Discount Amount. All of these 8 objects have been successfully updated with Special Discount Amount as the model item replacing Discount Amount. Note also that the data lineage has been checked and the reports validated at the same time. All 8 reports are now updated and ready to use.
In this example we only had 8 reporting objects that were identified and updated but it would work just as well if it were 80 or 800 reporting objects.
Note what we’ve done here: Not only have we been able to do an impact analysis of a model item that is changing, but as we have done that, we’ve been able to do a mass update to all of the impacted items that need to be changed. And finally, as part of the mass update itself, we have validated that all of the objects that have been updated will successfully run. A three part closed loop process:
- 1. Identify impact
- 2. Perform mass update
- 3. Validation
But this is only one model item that has been updated? What if we have multiple model items that need to be updated at the same time?
If you have multiple model items that need to be done these can all be included in one mass update as shown in the example below (Figure 3). You can simply upload all of the items from a comma delimited file and these can be made all at once and also validated at the same time insuring that the reporting objects using them will run without failure. Another use case example where this is helpful is if you are doing a refactoring of multiple items in a model.
Specification Impact Analysis
Specifications within Cognos Analytics can cover a broad range of metadata. A real world use case example of where specification impact analysis can be really useful is where you may be using a logo on reporting objects and that needs to be changed or replaced with an updated logo. Here we will do a search and replace of a logo (Figure 4).
The preview (Figure 5) shows that filtering on the package BI Reporting covered 61 objects but only turned up 2 objects that are using the current logo.
We are going execute this change and make the logo update (Figure 6). For this example we have intentionally chosen a report set that has very few reports with a logo that needs to be changed. In a real world environment this would likely be much larger and the ability to do mass updates to specifications could be a real time saver.
Another use case example for specification impact and change could be font size on reporting objects (Figure 7). You need only create a string copied from the documentation of a current reporting object and simply change the font size as shown below. There are many other examples where being able to change specifications on multiple items can be useful.
Database Impact Analysis
Database impact analysis is one of the most frequent analytical tasks performed by Cognos administrators. In this example we want to look across our entire Cognos environment and see which databases contain a table labeled Customer. We may then want to know if these all use the same definition of Customer, and if they are different, why they differ. We have the ability to search by database, table and column but here we will select just the table Customer (Figure 8).
This analysis (Figure 9) shows that there are a total 417 objects using the table labeled Customer across a total of 7 different databases. If we needed to at this point we could also add in more detail and get the actual named objects using this table for each database. But if our goal is to understand how many Customer tables we have and how they may differ we may have enough information here to get started.
Summary
All of these different examples are things that can be done quickly and easily. But beyond speed and simplicity is the knowledge that you’ve captured ALL of the objects of a given type that may be impacted across the entire Cognos Content Store. This assurance means that the impact analysis and its resulting changes are a one and done process. You don’t have to worry, “Did I get it all”?
So what’s the secret here? Two things:
- Having access to all of the Cognos metadata in way that you can use it to manage change in multiple different ways. The examples we’ve shown here cannot be done by trying to get data directly from the Content Store itself. Not only will you not be able to get the data you need, but what you do get will be very limited and incomplete. You need all of the Content Store metadata and be able to persist it for use in multiple different ways.
- It also requires the right tools to not only do the impact analysis itself, but to be able to make the changes to the objects identified as needing to be changed. This involves mass updates to objects. Sometimes just a few but it could involve hundreds or more. Plus, the right tool(s) can also validate that the changes have been correctly made thus insuring a closed loop system for managing change; from impact analysis to applying the fix to full validation of the change.
For more on this subject check out a related blog series on the Cognos Metadata Universe. With the right tools to use all of this metadata, impact analysis and the change that goes with it becomes a routine set of tasks to be done on whatever basis is needed.
Read more: Impact Analysis in Cognos Analytics, Part II
