Data Transformations

split(x, f, drop = FALSE, …)，x是待划分的向量或数据框，f是定义了组关系的因子变量。

library(MASS)
g <- split(Cars93$MPG.city, Cars93$Origin)
class(g)

## [1] "list"

names(g)

## [1] "USA"     "non-USA"

c(median(g[[1]]), median(g[[2]]))

## [1] 20 22

lapply(lst, fun)和sapply(lst, fun)，前者返回一个列表，后者优先返回向量，sapply里的s表示simplify。

Lst <- list(a = rnorm(100), b = rnorm(100), c = rnorm(100))
lapply(Lst, range)

## $a
## [1] -2.859  2.976
##
## $b
## [1] -1.961  2.906
##
## $c
## [1] -2.403  3.363

sapply(Lst, range)

##           a      b      c
## [1,] -2.859 -1.961 -2.403
## [2,]  2.976  2.906  3.363

对矩阵或数据框的行或列应用函数

1. 对矩阵行apply(mat, 1, fun)。
2. 对矩阵列apply(mat, 2, fun)。
3. 由于data frame是由其列为元素的列表，所以有lapply(dfrm, fun)和sapply(dfrm, fun)。
4. 假设resp是响应变量（response variable），pred是一个数据框，每列为一个predictor。cors <- sapply(pred, cor, y = resp)会计算pred的每列和y的相关系数。

resp <- rnorm(n = 10, mean = 0, sd = 1)
pred <- as.data.frame(matrix(rnorm(n = 10 * 100, mean = 0, sd = 1), 10, 100))
cors <- sapply(pred, cor, y = resp)
mask <- (rank(-abs(cors)) <= 10)  #函数rank给出从小到大的排序
best.pred <- pred[, mask]

sapply(Cars93, class)

##       Manufacturer              Model               Type
##           "factor"           "factor"           "factor"
##          Min.Price              Price          Max.Price
##          "numeric"          "numeric"          "numeric"
##           MPG.city        MPG.highway            AirBags
##          "integer"          "integer"           "factor"
##         DriveTrain          Cylinders         EngineSize
##           "factor"           "factor"          "numeric"
##         Horsepower                RPM       Rev.per.mile
##          "integer"          "integer"          "integer"
##    Man.trans.avail Fuel.tank.capacity         Passengers
##           "factor"          "numeric"          "integer"
##             Length          Wheelbase              Width
##          "integer"          "integer"          "integer"
##        Turn.circle     Rear.seat.room       Luggage.room
##          "integer"          "numeric"          "integer"
##             Weight             Origin               Make
##          "integer"           "factor"           "factor"

tapply：对数据向量按因子分组应用函数

对数据框的行按因子分组应用函数，by(dfrm, fact, fun)

非向量化函数的向量化，mapply(f, vec1, vec2, …, vecN)，f有N个参数。

mapply(rep, 1:4, 4:1)

## [[1]]
## [1] 1 1 1 1
##
## [[2]]
## [1] 2 2 2
##
## [[3]]
## [1] 3 3
##
## [[4]]
## [1] 4

mapply(rep, times = 1:4, x = 4:1)

## [[1]]
## [1] 4
##
## [[2]]
## [1] 3 3
##
## [[3]]
## [1] 2 2 2
##
## [[4]]
## [1] 1 1 1 1

gcd <- function(a, b) {if (b == 0) return(a) else return(gcd(b, a%%b))
}
mapply(gcd, c(1, 2, 3), c(9, 6, 3))

## [1] 1 2 3

参考文献

R cookbook 第六章