If there is one book that every beginning R user coming from a programming background should have, it is Spector’s Data Manipulation with R. New R users with analytic backgrounds and experience with software packages such as SAS and SPSS will do well to start with Muenchen’s R for SPSS and SAS users, especially given that a free abbreviated version is available, but those users should also make Data Manipulation with R a quick second addition to their library.
The text of this book is as concise and to the point as its title. It covers almost every relevant data manipulation topic in R, from modes and classes, through accessing data via database connections, to complex reshaping and aggregating functions. It has copious examples and the text hits just the right level of sophistication for the individual who has some experience with programming, but little experience with R idioms and data manipulation techniques.
My only critique of this book is that it skips over the basics of creating user-defined functions for data manipulation tasks. Spector addresses mapping functions to various data structures, but it seems likely that, at this level, the average R analyst would be better served by a discussion of how to simply create a function in R. Keep in mind that if you are looking for that type of information, you will need to look elsewhere. The same is true if you are looking for any sort of statistical instruction, as Data Manipulation with R focuses almost exclusively on programming.
Overall, I highly recommend this book. At around $45 USD, it is well worth the price. You’ll breeze through it on your first pass, but if you’re new to R you will get your money’s worth out of it as a reference text.