We have to be honest—R was not originally designed for handling great amounts of data. It was actually an instrument conceived for being employed within the academic world, where having fifty points of data, especially at that time, was some kind of miraculous event. From this admission comes the observation that R is not naturally provided with the right features for heavy data handling. Nevertheless, we can once again leverage here the great flexibility of our beloved language by means of a package developed to establish a communication between R and some of the most popular data warehouse solutions, such as those previously mentioned, MongoDB and Hadoop. Having these packages at your disposal, such as mongolite or rhdfs, will give you the possibility to leverage the related  Data Warehouse solution within your data mining project. In the hypothesis of being leveraging data already stored within your company, research lab or university, this will also mean to have the possibility to directly access them.