One of the frequent data structures I use in R is a data frame. Data frames are similar to matrices except they allow different types of variables in each column. I always rejoice when I can reduce the analysis at hand to a matrix of numbers. R is inherently faster at dealing with matrices than data frames. When I get data that is millions of rows long and over a hundred columns big, R still works, but slows down significantly. Lately, I spend most of my time figuring out the fastest way to transform data and perform calculations on big data frames.
Here is a great solution from the folks at R-studio. I had to post this to get the word out.
'dplyr' is a package that extends the plyr package to data frames and increases the speed tremendously. I plan to post examples using this new package soon.