Online courses like Coursera.org has a good selection for statistics using R. For example, this one from Princeton uses R as their teaching software. There are also intensive courses provided by private companies like this one.
Since you mentioned you have statistical background, I believe you also know at least one other statistical software package? If that's the case, I'd recommend first dig out a few comprehensive analyses you have done before, and try to replicate that with R. Through the process, you would need to consult the materials online and get more familiar with the language, and because you have your old output, you can get to compare your results with the model answers. In the process, remember to create and maintain a "cheat sheet" like this PDF document. It'd save a lot of time.
I have been using this language for about 8 years, started with S+. I use it almost everyday but to be frank, I never really endeavored to "learn" R. R is like a giant national park or theme park that you can just take a small chunk in each visit. There are basics like import/export, assigning object, writing function, etc. but then pretty much beyond linear mixed model it's everyone's own game. I hunt for the package that fits my need, and I study the package's manual. And all those basics are generously provided by users online.