By Paul Hausser, Envisn, Inc.
Our recent blogs on Cognos audit data have focused on getting it to where it can be used easily for anything you want to use it for. Once it’s been normalized this assumes that the data are consistent and capture everything within Cognos. Unfortunately this isn’t always true. So how do we address this?
There are two primary dimensions to the audit data; User and Object (or content).
One would expect that every audit data record can be tied to both of these dimensions but that isn’t always the case for one reason or another. In the audit data record reconciliation shown below for the month of July for a mid to large Cognos environment, we have a total record count in our fact table of 42,557.
This number is a control total. It represents the total number of audit data records in our fact table because each one has an audit data action within Cognos. In many cases the Cognos audit data action record may not have sufficient information to determine object type or user. UniVisn’s unique process uses advanced algorithms to compensate for this and increase the match rate wherever possible.
In our records reconciliation we want to see how many non-matches we have on both object type and user. The No Content Ref. Count of 50 reflects those records we cannot identify an object type within the content dimension.
In the lower section of this table No User Ref Count of 35 reflects those records where no user match exists.
The good news is that both of these numbers combined represent two tenths of one percent (0.2%) of the total number of records 42,557. Said differently, we’re able to reconcile both user and object dimension record count at 99.8% and utilize virtually all of the audit data records for the month of July.
These numbers are well within the parameters of accuracy for any purposes of reporting. Plus, we have the record detail for both counts if we wanted to research why user or object type may be some missing data.
Having a system like this that not only maximizes audit record match around the key dimensions but also automatically generates key reconciling numbers for every period processed significantly increases the confidence of users that the audit data numbers are fact based. This is critical for insuring that they will actually be used and trusted in the reporting process.
