I'm currently in the process of assisting with the integrating of Informatica with Qlikview using the new Informatica connector to generate QVX files.
The massive benefit of using Informatica is that it gives you confidence in the data. Within a large organisation there are likely to be multiple datasources (straight db's and warehouses) that are used by more than one reporting tool - Qlikview, BO, Cognos etc and by using Informatica we can have a single extract and transformation layer instead of several versions all with their own differences.
That doesn't sound a big deal (and in a small deployment it wouldn't be) but think of all the ET(L) to be maintained across say a bank or insurance company; it's so much more reliable to use a single extract process that results in one set of code to maintain. In addition Informatica provides great data lineage, meaning your MI is fully auditable ie: you know exactly where field 'X' in a Qlikview dashboard has come from and how it's been transformed.
It also allows Qlikview (through the QVX files) to benefit from ETL already generated in Informatica; there's no need to re-write existing complex transforms that in some cases would be totally impractical in QV script.
For me it goes to the heart of Qlikview's sales pitch 'one version of the truth' - with multiple ETL processes how can you possibly be sure that the picture you present in Qlikview is the same as the one being presented in BO; you can't - plus when it comes to audit (SOX etc) you can quickly and easily prove that the MI you're presenting is correct and reliable.
Due to the cost of Informatica I can only ever see it of being of use for larger organisations and those with a stringent audit requirement.
So far it looks as though the Connector is a great tool; v1.01 has vastly improved the speed to make it a viable solution.
Hope that clears things up.
All the best,
Matt - Visual Analytics Ltd
.png "Luminary Alumni"){kind=icon}
