learnR_basicdataprocessing_5

learnR_basicdataprocessing_5

gender <- c('F','F','M',NA,'F','M')
age <- c(12,56,34,67,NA,23)
student <-c(NA,'Tom','Toby','Jreey','Curry','Allen')
score <-c(99,NA,60,75,65,85)
leader <- data.frame(gender,age,student,score,stringsAsFactors=FALSE)

1. 变量操作
1.1 创建新变量
variable <- expression

Example1:
mydata <- data.frame(x1 = c(1,2,3,4), x2 = c(5,6,7,8))
mydata$sumx <- mydata$x1 + mydata$x2
mydata$meanx <- (mydata$x1 + mydata$x2) / 2

Example2:
attach(mydata)
mydata$sumx <- x1 + x2
mydata$meanx <- (x1 + x2) / 2
detach(mydata)

Example3: #transform
mydata <- transform(mydata, sumx = x1 + x2, meanx = (x1 + x2) / 2)

1.2 变量重编码
variable[condition] <- expression

1.2.1 
Example1:
df <- data.frame(age=c(20,45,80))
df <- within(df,{
            agecat <- NA
            agecat[age > 0 & age < 30] <- 'Young'
            agecat[age >= 30 & age <= 50] <- 'Middle Aged'
            agecat[age > 50] <- 'Elder'
})

#重编码函数
1.2.2 recode() #Recode a Variable

example2:
library(car)
x <- rep(1:3, 3)
recode(x, "c(1,2) = 'A'; else = 'B'")
recode(x, "1:2 = 'C'; 3 = 'D'" )

1.2.3 recodeVar() #Recode values of vector

example3:
library(doBy)
x <- c(NA, rep(1:3,3), NA)
recodeVar(x, src=list(c(1,2)), tgt=list('A'), default='dft')
recodeVar(x, src=list(c(1,2)), tgt=list('B'), default='dft',keep.na=F)

1.2.4 cut() #Convert Numeric to Factor

example4:
aaa <- c(1:5, 2:7)
cut(aaa, 3)
cut(aaa, 3, dig.lab=4, ordered=T)

1.3 变量重命名
df <- data.frame(name=c('aa','bb','cc'),age=c(10,NA,110))

1.3.1 fix(df) #交互式修改

1.3.2 rename()
Example:
library(reshape)
df <- rename(df, c(name='new_name', age='new_age'))

1.3.3 names()
names(df) <- c('n_name','n_age')

1.4 缺失值
NA：占位缺失值
NaN：占位错误值
NULL：空集
Inf：占位无穷大
判断 ：is.na(df)
噪声数据转换为NA：df$age[df$age == 110] <- NA

排除NA：
eg1: y <- sum(x=c(1,2,NA,3), na.rm=TRUE) #实际是当作0处理
eg2: new_df <- na.omit(df) #删除NA存在的行

1.5 日期值
Sys.Date()      #年月日
date()          #具体年月日，星期，时钟时间

#日期默认格式：yyyy-mm-dd
my_dates <- as.Date(c('2017/03/22', '2017/03/25'))
her_dates <- as.Date('03/22/2017', '%m/%d/%Y')

1.5.1 format() #提取日期特定部分
eg1:
today <- Sys.Date() #date()貌似不好提取特定部分
today_week <- format(today, format='%A')

eg2: #提取某日的星期数
birthday <- as.Date('1999-11-25')
one_week <- format(birthday, format='%A')

1.5.2 difftime() #日期差值
eg1:
difftime(today, birthday, units='weeks')
eg2:
days <- today - birthday

1.5.3 其他日期函数
#连接日期和时钟时间
strftime()      #class='POSIXlt POSIXt'
strptime()      #class='chartacter'
eg:
p_time <- strptime('25/3/17 20:01:30.007', '%d/%m/%y %H:%M:%OS')#'POSIXlt POSIXt'
f_time <- strftime('25/3/17 20:01:30.007', format='%d/%m/%y %H:%M:%OS')#'character'
is_s <- '25/3/17 20:01:30.007'
f_time == is_s  #FALSE

#日期转换字符串
str_date <- as.character(date)

other:
help(ISOdatetime)
library(lubridate)
library(fCalendar)

1.6 类型转换
as.numeric()        #is.numeric()
as.character()
as.vector()
as.matrix()
as.data.frame()
as.factor()
as.logical()
结合控制流（if-then）使用时，功能强大。

1.7 排序 
order()
eg:
gender <- c('F','F','M',NA,'F','M')
age <- c(12,56,34,67,NA,23)
student <-c(NA,'Tom','Toby','Jreey','Curry','Allen')
score <-c(99,NA,60,75,65,85)
leader <- data.frame(gender,age,student,score,stringsAsFactors=FALSE)
new_data1 <- leader[order(gender, -age, na.last=TRUE),] #逆序/leader$gender

1.8 数据集的合并
1.8.1 添加列
eg1:    #有共同索引
total_1 <- merge(df_A, df_B, by=c('ID','Postcode'))
eg2:    #无需共同索引
total_2 <- cbind(A, B) #A，B必须有相同的行数和顺序

1.8.2 添加行
eg3:    #两个数据框必须有相同的变量，顺序不做要求。
total_3 <- rbind(df_A, df_B)

1.9 数据集取子集
1.9.1 保留变量（这里特指‘列’）
new_df <- df[, column_index]

eg1:
df1 <- df[,c(2:4)] #保留某些列，即所谓的变量

eg2:
my_df <- data.frame(matrix(1:12,3,4))
vars1 <- paste('X', 1:4, sep='')
new_df <- my_df[vars1]

1.9.2 剔除变量（这里特指‘列’）

#1 new_df <- df[!vars]
vars2 <- names(my_df) %in% c('X3','X4')
new_df <-my_df[!vars2]

#2 new_df <- df[c(-var3,-var4)]
new_df <- my_df[c(-3,-4)]

#最后，相同的变量删除工作亦可通过：
my_df$X3 <- my_df$X4 <- NULL

1.9.3 保留观测（相对的，这里特指‘行’）
new_df <- df[row_index,]

eg1:
df1 <- df[c(2:4),]

eg2:
new_data2 <- leader[which(gender == 'M' | age <= 34),]#一般leader$gender少出错

1.9.4 subset()函数
new_data3 <- subset(leader,gender=='M' & age<=34,select=c(age:score))#select

1.9.5 随机抽样
my_sample <- leader[sample(1:nrow(leader), 3, replace = FALSE),]#无放回随机
library(sampling)
library(survey)

1.10 使用SQL语句
eg:
library(sqldf)
newdf <- sqldf('select * from mtcars where carb=1 order by mpg', row.names=TRUE)
sqldf('select avg(mpg) as avg_mpg, avg(disp) as avg_disp, gear from mtcars 
        where cyl in (4,6) group by gear')