Hi,
we have two new releases, as you may have guessed from the subject. rmr2 3.2.0 is mostly a bugfix release with just one little feature added, which forced us to bump the minor number. plyrmr 0.4.0 introduces fast aggregation by integrating with dplyr a bit more. You may find this evolution a little bit of a work-in-progress but I wanted to get something out to hear from you. The idea is that when you group by a set of columns that defines many small groups, programs become inefficient. dplyr has attacked, if not solved, that problem for the most common aggregations. So we tried to bring into plyrmr some of that goodness with the least amount of overhead. The other improvement is that, like rmr2, plyrmr now has a system of package options and allows multiple backends. Two backends, local and hadoop, are inherited from rmr2. More to come in the future, stay tuned. Feedback and comments welcome as usual. Thanks
Antonio