ETL for a data warehouse? If this warehouse has more than a few tables then it's clearly DI Studio.
That's the specialised SAS tool to build and maintain ETL jobs.
I agree 150% with Shane's statements - especially when it comes to DIS 4.2.
Purchasing DIS might cause some initial cost both in license fees and training costs. But once developers are up to speed this investment is paid back quite fast in shorter and more stable development cycles and in easier maintainable jobs.
The current project I'm working in includes a big ETL implementation with DIS 4.2. All developers are on senior level but I'm convinced we wouldn't have stood a chance to meet the deadlines using EG instead of DIS for our ETL's.
I think the answer needs to be in perspective. It really depends on what kind of "data warehouse" we are talking about.
For an enterprise level, true data warehouse, there's no question that a proper ETL is a must.
For a small "data warehouse" with 1 or 2 star schema and 1-2 fact tables with a few dimensions, where no table is bigger than a few thousand records, it's not even a "data warehouse" per se. On top of that, if we only have about a few hundred records a day to deal with, I would think implementing a full-blown ETL tool is an major overkill.
SAS EG should be more than sufficient - maybe it's already an overkill because for that simple "data warehouse", one can just write a few SQL codes to achieve all that "ETL". I would think the volume of the data coupled with the complexity of the data and the transformation should be the deciding factors.