一言不合就学R

商务大数据智能分析 之 R2

Lecturer : 申 旌 周
jingzhou_shen@stu.xjtu.edu.cn

Instructor : 常 象 宇
xiangyuchang@xjtu.edu.cn

2017年3月12日

• 基础运算
• 数据类型
• 数据结构
• 数据读写
• 控制流

• 基础运算
• 数据类型
• 数据结构
• 数据读写
• 控制流

• 逻辑运算
  &,|,!
  与或非

22 %% 6

[1] 4

floor(22/6)

[1] 3

sqrt(22/6)

[1] 1.914854

log(22/6)

[1] 1.299283

• <-
• =
• <<-

• <- 双向赋值

asg <- 2

asg

[1] 2

3 -> asg

asg

[1] 3

• = 单向传值

asg = 1

asg

[1] 1

# 这是错误的写法
1 = asg  

• = 函数参数传值

Usage
matrix(data = NA, nrow = 1, ncol = 1, ...)

matrix(1:6, 3)

     [,1] [,2]
[1,]    1    4
[2,]    2    5
[3,]    3    6

• = 函数参数传值

GOOD

matrix(1:6, ncol = 3)

     [,1] [,2] [,3]
[1,]    1    3    5
[2,]    2    4    6

BAD

matrix(1:6, ncol <- 3)

     [,1] [,2]
[1,]    1    4
[2,]    2    5
[3,]    3    6

• = 函数参数传值

BAD函数参数传值(续)

ls() # 显示当前环境变量

[1] "asg"  "ncol"

mean(x = 1:3)

[1] 2

• = 函数参数传值

BAD函数参数传值(续)

ls()

[1] "asg"  "ncol"

mean(x <- 1:3)

[1] 2

ls()

[1] "asg"  "ncol" "x"   

• <- 与 = 小结

除函数参数传值外, = 可被替换为 <- , 反之则不然, 如: system.time(A <- matrix(1:6, 3))中只能用 <-

• <<- 表示给上一层环境中的变量赋值

• 建议：进行赋值运算时，使用 <- 而非 = .

  • Google's R Style Guide

• 基础运算
• 数据类型
• 数据结构
• 数据读写
• 控制流

原子型对象

• 逻辑型：logical
• 字符型：character
• 整数型：integer
• 实数型：numeric (double)
• 复数型：complex

is.logical(T)

[1] TRUE

is.logical(1)

[1] FALSE

原子型对象

• 逻辑型：logical
• 字符型：character
• 整数型：integer
• 实数型：numeric (double)
• 复数型：complex

typeof(7)

[1] "double"

is.numeric(7)

[1] TRUE

原子型对象

• 逻辑型：logical
• 字符型：character
• 整数型：integer
• 实数型：numeric (double)
• 复数型：complex

is.integer(7)

[1] FALSE

is.character("K")

[1] TRUE

• 基础运算
• 数据类型
• 数据结构
• 数据读写
• 控制流

非原子型对象

• 向量：Vectors
• 矩阵：Matrices
• 数据框：Data Frames
• 列表：Lists
• 因子：Factors

非原子型对象

• 向量：Vectors
• 矩阵：Matrices
• 数据框：Data Frames
• 列表：Lists

vector(mode = "character", length = 3)

[1] "" "" ""

v <- c(0, 1, -2)

typeof(v)

[1] "double"

class(v)

[1] "numeric"

length(v)

[1] 3

as.character(v)

[1] "0"  "1"  "-2"

as.logical(v)

[1] FALSE  TRUE  TRUE

class(v)

[1] "numeric"

class(as.logical(v))

[1] "logical"

v[c(2, 3)]

[1]  1 -2

v[2:3]

[1]  1 -2

v[-2]

[1]  0 -2

v[which(v < 1)]

[1]  0 -2

names(v)

NULL

names(v) <- c("张三", "李四", "王五")

names(v)

[1] "张三" "李四" "王五"

v[c("王五", "李四")]

王五 李四 
  -2    1 

t <- c(2, 1, 3)

(v + t)/2

张三 李四 王五 
 1.0  1.0  0.5 

t > v 

 张三  李四  王五 
 TRUE FALSE  TRUE 

summary(v)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
-2.0000 -1.0000  0.0000 -0.3333  0.5000  1.0000 

非原子型对象

• 向量：Vectors
• 矩阵：Matrices
• 数据框：Data Frames
• 列表：Lists

M <- matrix(1:9, nrow = 3, ncol = 3)
print(M)

     [,1] [,2] [,3]
[1,]    1    4    7
[2,]    2    5    8
[3,]    3    6    9

length(M)

[1] 9

dim(M)

[1] 3 3

cbind(1:3, 4:6)

     [,1] [,2]
[1,]    1    4
[2,]    2    5
[3,]    3    6

rbind(1:3, 4:6)

     [,1] [,2] [,3]
[1,]    1    2    3
[2,]    4    5    6

attributes(M)

$dim
[1] 3 3

vec <- 1:9

vec

[1] 1 2 3 4 5 6 7 8 9

attributes(vec)

NULL

dim(vec) <- c(3, 3)

vec

     [,1] [,2] [,3]
[1,]    1    4    7
[2,]    2    5    8
[3,]    3    6    9

M[2,3]

[1] 8

M[1, ]

[1] 1 4 7

M[, 2]

[1] 4 5 6

Q <- matrix(rep(10, 9), 3, 3)

M * Q

     [,1] [,2] [,3]
[1,]   10   40   70
[2,]   20   50   80
[3,]   30   60   90

M %*% Q

     [,1] [,2] [,3]
[1,]  120  120  120
[2,]  150  150  150
[3,]  180  180  180

colMeans(M)

[1] 2 5 8

colSums(M)

[1]  6 15 24

colnames(M)

NULL

rownames(M)

NULL

非原子型对象

• 向量：Vectors
• 矩阵：Matrices
• 数据框：Data Frames
• 列表：Lists

nutrition <- data.frame(fruit = c("apples", "pears", "bananas"), calories = c(52, 57, 92))

head(nutrition, 3)

    fruit calories
1  apples       52
2   pears       57
3 bananas       92

ncol(nutrition)

[1] 2

nrow(nutrition)

[1] 3

names(nutrition)[1] <- "type"

attributes(nutrition)

$names
[1] "type"     "calories"

$row.names
[1] 1 2 3

$class
[1] "data.frame"

colnames(nutrition)

[1] "type"     "calories"

cbind(nutrition, nutrition)

     type calories    type calories
1  apples       52  apples       52
2   pears       57   pears       57
3 bananas       92 bananas       92

rbind(nutrition, nutrition)

     type calories
1  apples       52
2   pears       57
3 bananas       92
4  apples       52
5   pears       57
6 bananas       92

str(nutrition)

'data.frame':   3 obs. of  2 variables:
 $ type    : Factor w/ 3 levels "apples","bananas",..: 1 3 2
 $ calories: num  52 57 92

summary(nutrition)

      type      calories   
 apples :1   Min.   :52.0  
 bananas:1   1st Qu.:54.5  
 pears  :1   Median :57.0  
             Mean   :67.0  
             3rd Qu.:74.5  
             Max.   :92.0  

nutrition[which(nutrition$type == "apples"), ]

    type calories
1 apples       52

nutrition$calories # 返回结果为向量 

[1] 52 57 92

nutrition[["calories"]] # 返回结果为向量

[1] 52 57 92

nutrition["calories"] # 返回结果为数据框

  calories
1       52
2       57
3       92

calorie <- nutrition[["calories"]]

max(calorie)

[1] 92

subset(nutrition, select = "type", subset = calories > 80)

     type
3 bananas

row.names(nutrition)

[1] "1" "2" "3"

rownames(nutrition)

[1] "1" "2" "3"

非原子型对象

• 向量：Vectors
• 矩阵：Matrices
• 数据框：Data Frames
• 列表：Lists

list(
  c("张三", "李四", "王五", "刘六"), 
  c(183, 168, 177, 167), 
  c("男", "女", "男"))

[[1]]
[1] "张三" "李四" "王五" "刘六"

[[2]]
[1] 183 168 177 167

[[3]]
[1] "男" "女" "男"

L <- list(
  name = c("张三", "李四", "王五", "刘六"), 
  height = c(183, 168, 177, 167), 
  gender = c("男", "女", "男"))

L

$name
[1] "张三" "李四" "王五" "刘六"

$height
[1] 183 168 177 167

$gender
[1] "男" "女" "男"

str(L)

List of 3
 $ name  : chr [1:4] "张三" "李四" "王五" "刘六"
 $ height: num [1:4] 183 168 177 167
 $ gender: chr [1:3] "男" "女" "男"

Doraemon.bag <- list(
  char = "K",
  num = c(2, 2, 6),
  mat = matrix(6, 2, 2),
  df = data.frame(A = 1:3, I = 4:6),
  list = list(D = 1, M = 9)
)

str(Doraemon.bag)

List of 5
 $ char: chr "K"
 $ num : num [1:3] 2 2 6
 $ mat : num [1:2, 1:2] 6 6 6 6
 $ df  :'data.frame':   3 obs. of  2 variables:
  ..$ A: int [1:3] 1 2 3
  ..$ I: int [1:3] 4 5 6
 $ list:List of 2
  ..$ D: num 1
  ..$ M: num 9

L$name

[1] "张三" "李四" "王五" "刘六"

L[["name"]]

[1] "张三" "李四" "王五" "刘六"

L["name"]

$name
[1] "张三" "李四" "王五" "刘六"

c(is.vector(L$name), is.vector(L[["name"]]),   
  is.list(L["name"]))

[1] TRUE TRUE TRUE

f <- factor(c("yes", "yes", "no", "no", "yes"))

f

[1] yes yes no  no  yes
Levels: no yes

attributes(f)

$levels
[1] "no"  "yes"

$class
[1] "factor"

f <- factor(
  c("yes", "yes", "no", "no", "yes"),    
  levels = c("yes", "no"))

f

[1] yes yes no  no  yes
Levels: yes no

attr(f, "levels") <- c("y", "n")

unclass(f)

[1] 1 1 2 2 1
attr(,"levels")
[1] "y" "n"

缺失值

• NaN
• NA

q <- c(1, NaN, NA, 4)

class(q)

[1] "numeric"

is.nan(q)

[1] FALSE  TRUE FALSE FALSE

is.na(q)

[1] FALSE  TRUE  TRUE FALSE

缺失值

• NaN
• NA

mean(q)

[1] NaN

mean(q, na.rm = T)

[1] 2.5

1. 均为R中的对象
2. R中的对象可能具有以下属性
   • names, dimnames
   • dimensions
   • class
   • length
   • …
3. 对象的属性可由函数 attributes() 访问

• 基础运算
• 数据类型
• 数据结构
• 数据读写
• 控制流

• 读取csv

data <- read.csv("./data/data.csv", header = TRUE) 

data

  姓名 性别 年龄
1 张三   女   11
2 李四   女   12
3 王五   男   12

class(data)

[1] "data.frame"

• 写入csv

write.csv(data, file = "./data/dataout.csv")

• 基础运算
• 数据类型
• 数据结构
• 数据读写
• 控制流

• if, else
• for and something new
• while
• repeat
• break
• next
• return

s <- 1:5
ifelse(s > 3, yes = "Good", no = "Fair")

[1] "Fair" "Fair" "Fair" "Good" "Good"

a <- 1

if(a > 2){
  b <- 1
}else{
  b <- 0
}

b

[1] 0

a <- -1

if(a > 2){
  b <- 1
}else if(a < 0){
  b <- 1
}else{
  b <- 0
}

b

[1] 1

for(i in 1:3){
  print(i) 
}

[1] 1
[1] 2
[1] 3

M <- matrix(1:2, 2, 3)

M

     [,1] [,2] [,3]
[1,]    1    1    1
[2,]    2    2    2

• 对列求和

numeric(3)

[1] 0 0 0

col.sum <- numeric(ncol(M))
for(j in 1:ncol(M)){
  for(i in 1:nrow(M)){
    col.sum[j] <- 
      col.sum[j] + M[i, j]
  }
}

col.sum

[1] 3 3 3

• 对列求和

M

     [,1] [,2] [,3]
[1,]    1    1    1
[2,]    2    2    2

• 向量化运算

col.sum <- numeric(ncol(M))
for(i in 1:nrow(M)){
  col.sum <- col.sum + M[i, ]
}

col.sum

[1] 3 3 3

apply(M, 2, sum)

[1] 3 3 3

apply(M, 1, sum)

[1] 3 6

L <- list(
  a = matrix(1:3, 1), 
  b = data.frame(
    char = c("A", "B"), 
    logic = c(T, F)), 
  c = list(
    1:3, 
    "T",
    matrix(1:6, 2, 3)))

L 

$a
     [,1] [,2] [,3]
[1,]    1    2    3

$b
  char logic
1    A  TRUE
2    B FALSE

$c
$c[[1]]
[1] 1 2 3

$c[[2]]
[1] "T"

$c[[3]]
     [,1] [,2] [,3]
[1,]    1    3    5
[2,]    2    4    6

lapply(L, class)

$a
[1] "matrix"

$b
[1] "data.frame"

$c
[1] "list"

sapply(L, class)

           a            b            c 
    "matrix" "data.frame"       "list" 

mapply(rep, 1:6, 6:1)

[[1]]
[1] 1 1 1 1 1 1

[[2]]
[1] 2 2 2 2 2

[[3]]
[1] 3 3 3 3

[[4]]
[1] 4 4 4

[[5]]
[1] 5 5

[[6]]
[1] 6

set.seed(6)
n <- 9
bike <- data.frame(
  orgn = sample(c("SE", "S", "N"), n, replace = T),
  dest = sample(c("SE", "S", "N"), n, replace = T),
  flux = round(rnorm(n, 18))) 

bike

  orgn dest flux
1    S   SE   17
2    N    S   20
3   SE    N   17
4    S   SE   19
5    N   SE   18
6    N    N   17
7    N   SE   18
8    N    S   20
9    S    N   17

table(bike$orgn, bike$dest) 

     N S SE
  N  1 2  2
  S  1 0  2
  SE 1 0  0

tapply(
  bike$flux, 
  bike[c("orgn", "dest")], 
  mean)

    dest
orgn  N  S SE
  N  17 20 18
  S  17 NA 18
  SE 17 NA NA

• next

for(i in 1:10){
  if(i <= 7){ 
    next
  }
  print(i)
  i <- i + 1
}

[1] 8
[1] 9
[1] 10

• return

square <- function(x) {
  return(x^2)
}

square(11)

[1] 121

square <- function(x) {
  x^2
}

square(11)

[1] 121