I've been having alot of luck using R for aggregation on small (less than a gig) data sets. For aggregation work, time series analysis, I'm starting to just use R.
I'm enrolled in the Data Science program at the University of Washington this fall so I needed to get used to R (and I already know Hadoop basics).
I have some code to shell out from .Net, works well so far.