R語言函數式編程purrr的應用

01-28

S語言的設計者、R語言核心團隊成員Chambers（2016）在Extending R一書中提到R語言中的三大理念。

Everything that exists in R is an object. OBJECT
Everything that happens in R is a function call.FUNCTION
Interfaces to other software are part of R. INTERFACE

在數據分析中，R語言函數式編程用的比較多。為此，Wickham等人編寫一個高級函數編程語言包purrr。

purrr enhances Rs functional programming (FP) toolkit by providing a complete and consistent set of tools for working with functions and vectors.

purrr在0.2.2版本之後有新的變化，變得更輕，更底層，分解為purrr和新的purrrlyr包。

purrr no longer depends on dplyr, lazyeval, or Rcpp. This makes the dependency graph of the tidyverse simpler, and makes purrr more suitable as a dependency of lower-level packages. This means that data-frame based mappers (dmap(), dmap_at(), dmap_if(), invoke_rows(), slice_rows(), map_rows(), by_slice(), by_row(), and unslice()) have been moved to a new package, purrrlyr. This is a bit of an aggressive change but it needed to be done, and in this case I think its better to rip the band aid off quickly.

R語言在運行時候速度確實一般，因此很多小夥伴得到的忠告就是避免使用循環，而是採用apply系列函數。（其實，在必要的時候，使用循環還是很方便的。）

apply系列函數簡單應用

應用場景，求均值。分別採用了apply(mtcars,2,mean)、lapply(mtcars,mean)和sapply(mtcars,mean)。

> apply(mtcars,2,mean)n mpg cyl disp hp drat wt qsec vs am n 20.090625 6.187500 230.721875 146.687500 3.596563 3.217250 17.848750 0.437500 0.406250 n gear carb n 3.687500 2.812500 n> lapply(mtcars,mean)n$mpgn[1] 20.09062n$cyln[1] 6.1875n$dispn[1] 230.7219n$hpn[1] 146.6875n$dratn[1] 3.596563n$wtn[1] 3.21725n$qsecn[1] 17.84875n$vsn[1] 0.4375n$amn[1] 0.40625n$gearn[1] 3.6875n$carbn[1] 2.8125nn> sapply(mtcars,mean)n mpg cyl disp hp drat wt qsec vs am n 20.090625 6.187500 230.721875 146.687500 3.596563 3.217250 17.848750 0.437500 0.406250 n gear carb n 3.687500 2.812500 n

其實，這個時候採用函數式編程中常見的map也可以實現，上述的分析需求。R語言中的map需要大寫Map(函數,數據)。

> Map(mean,mtcars)n$mpgn[1] 20.09062n$cyln[1] 6.1875n$dispn[1] 230.7219n$hpn[1] 146.6875n$dratn[1] 3.596563n$wtn[1] 3.21725n$qsecn[1] 17.84875n$vsn[1] 0.4375n$amn[1] 0.40625n$gearn[1] 3.6875n$carbn[1] 2.8125n