I think that "R in Action" is the best "general-purpose" R book on the market. Previously, I would have recommended Joe Adler's "R in Nutshell" as the closest thing to an "R Bible", and Paul Teetor's "R Cookbook" as the beginner-friendly introduction. Now, "R in Action" comes close to giving one the best of both worlds, combining the accessibility of Teetor's book with the wide scope and thoroughness of Adler's. In adversarial terms, "R Cookbook", competing in the introduction-to-R pageant, charms the audience, but loses to "R in Action" on points - while "R in Nutshell", defending its hold on the "intermediate" sector, retains 100% support of multiple special-interest groups (including Machine Learning Alliance, Bioinformatics Computing League and Association for Advancement of Trellis Plots), but loses the median voter. It's a good idea to get two books, anyway, and this one deserves a spot on the list.
PS. Each of the three books splits the material into (a) "R fundamentals", including syntax, data structures and I/O, basic statistics and graphics, and (b) "Specialized statistics", starting from the linear regression. On (b), "R in Action" wins over "R in Nutshell" on depth, and distinguishes itself on (a), by offering valuable advice absent in "R in Nutshell". (Examples include importing data from Excel, SAS or SPSS, reading XML, parsing date strings, using string functions (!), reshaping data frames with "reshape" package, and - #1 on my list - SQL-querying data frames with "sqldf").
PPS. The one-star reviewer has quite a track record of drive-by poop-flinging - and if he did buy the "utterly useless" book (to burn it, I guess) directly from Manning, he paid $15-25 over Amazon's price.
PPPS. When I say "general-purpose", I primarily mean "not statistical". The latter group is itself split into generalist and specialized books; the main contenders in the first category seem to be (a) Crawley and (b) Maindonald and Brown, and out of these two, I would go for Maindonald and Brown.
PPPPS. Competition is catching up... "The R Primer" by Ekstrom (published in August 2011) looks quite nice, and beats "R in Action" on graphics coverage.