利用R语言的dplyr包进行数据转换

library(tidyverse)
library(nycflights13)    #利用该包中的flights数据

flights

#### R语言中的变量类型
# int——整数型变量
# dbl——双精度浮点数型变量，或称实数
# chr——字符串
# dttm——日期时间型变量
# lgl——逻辑型变量
# fctr——因子，即具有固定数目的值的分类变量
# date——日期型变量



####使用filter()筛选行
(jan_1 <- filter(flights,month == 1,day == 1))
#比较浮点数是否相等时，不能使用==，而应该使用near()
near(sqrt(2) ^ 2, 2)
#逻辑运算符或
filter(flights,month == 11 | month == 12)
#等价于
(nov_dec <- filter(flights,month %in% c(11,12)))
#判定一个值是否为缺失值
is.na(3)
#filter()默认排除条件为FALSE和NA的行。如果想保留缺失值，可以明确指出
df <- tibble(x = c(1, NA, 3))
filter(df, is.na(x) | x > 1)



####使用arrange()排列行
#按默认升序
arrange(flights,year,month,day)
#按降序
arrange(flights,desc(arr_delay))
#缺失值总是排在最后
df <- tibble(x = c(5,2,NA))
arrange(df,x)



####使用select()选择列
#按名称选择列
select(flights,year,month,day)
#选择year和day之间所有的列(包括year和day)
select(flights,year:day)
#选择不在year和day之间所有的列(不包括year和day)
select(flights,-(year:day))
#重命名变量
rename(flights,tail_num = tailnum)
#将几个变量移动到数据框的开头
select(flights,time_hour,air_time,everything())



####使用mutate()添加变量
#新增变量会放在最后
#创建一个新的更狭窄的数据集，以便能够看到新变量
flights_sml <- select(flights,
  year:day,
  ends_with("delay"),
  distance,
  air_time
)
flights_sml
mutate(flights_sml,
       gain = arr_delay - dep_delay,
       speed = distance / air_time * 60)
#创建的新列可以立即使用
mutate(flights_sml,
       gain = arr_delay - dep_delay,
       hours = air_time / 60,
       gain_per_hour = gain / hours)
#若只想保留新变量，使用transmute()函数
transmute(flights_sml,
          gain = arr_delay - dep_delay,
          hours = air_time / 60,
          gain_per_hour = gain / hours)
#可以与mutate()一起使用的向量化函数
# 算数运算符：+、-、*、/、^
# 模运算符：%/%(整除)和%%(求余)
transmute(flights,
          dep_time,
          hour = dep_time %/% 100,
          minute = dep_time %% 100
)
# 对数运算符：log()、log2()和log10()
# 偏移函数：lead()和lag()
(x <- 1:10)
lead(x)
lag(x)
# 累加和滚动聚合
# 累加和cumsum()
cumsum(x)
# 累加积cumprod()
cumprod(x)
# 累加最小值cummin()
cummin(x)
# 累加最大值cummax()
cummax(x)
# 逻辑比较：<、<=、>、>=、!=
# 排秩
y <- c(1,2,2,NA,3,4)
min_rank(y)
row_number(y)



####使用summarize()进行分组摘要
by_day <- group_by(flights, year, month, day)
summarize(by_day,delay = mean(dep_delay,na.rm = TRUE))


###使用管道组合多种操作
#研究每个目的地的距离和平均延误时间之间的关系
#普通写法
by_dest <- group_by(flights,dest)
delay <- summarize(by_dest,
                   count = n(),
                   dist = mean(distance,na.rm = TRUE),
                   delay = mean(arr_delay,na.rm = TRUE))
delay <- filter(delay, count > 20, dest != "HNL")
ggplot(data = delay, mapping = aes(x = dist, y = delay)) + 
  geom_point(aes(size = count), alpha = 1/3) + 
  geom_smooth(se = FALSE)


###管道写法
#管道的重点在于转换的过程而不是对象
delays <- flights %>%
  group_by(dest) %>%
  summarize(
    count = n(),
    dist = mean(distance, na.rm = TRUE), delay = mean(arr_delay, na.rm = TRUE)
  ) %>%
  filter(count > 20, dest != "HNL")


###缺失值
flights %>%
  group_by(year, month, day) %>%
  summarize(mean = mean(dep_delay, na.rm = TRUE))
#也可以先去除缺失值再进行分组摘要
not_cancelled <- flights %>%
  filter(!is.na(dep_delay), !is.na(arr_delay))
not_cancelled %>%
  group_by(year, month, day) %>%
  summarize(mean = mean(dep_delay))


###计数
#可以确定结论不是基于非常少的数据得出来的
delays <- not_cancelled %>%
  group_by(tailnum) %>%
  summarize(
    delay = mean(arr_delay)
  )
ggplot(delays, aes(x = delay)) + 
  geom_freqpoly(binwidth = 10)
#绘制航班数量和平均延误时间的散点图
delays <- not_cancelled %>%
  group_by(tailnum) %>%
  summarize(
    delay = mean(arr_delay,na.rm = TRUE),
    n = n()
  )
ggplot(data = delays, mapping = aes(x = n, y = delay)) +
  geom_point(alpha = 1/10)
#剔除观测值较少的飞机
delays %>%
  filter(n > 25) %>%
  ggplot(mapping = aes(x = n, y = delay)) +
  geom_point(alpha = 1/10)

##利用Lahman包中的数据计算大联盟中每个棒球队员的打击率（安打数/打数）
(batting <- as_tibble(Lahman::Batting))
batters <- batting %>%
  group_by(playerID) %>%
  summarize(
    ba = sum(H, na.rm = TRUE) / sum(AB, na.rm = TRUE),    #能力——ba
    ab = sum(AB, na.rm = TRUE)                            #击球数量——ab
  )
batters %>%
  filter(ab > 100) %>%
  ggplot(mapping = aes(x = ab, y = ba)) +              #在管道中不必再声明data
  geom_point() +
  geom_smooth(se = FALSE)


### 常用的摘要函数
##位置度量：mean()、median()
##分散程度度量：sd()、IQR(四分位距)、mad(绝对中位差)
##秩的度量：min()、quantile(x,0.25)、max()
##定位度量：first()、nth(x,2)、last()
##计数：n()、n_distinct(计算唯一值的数量)、count(仅用于计数)
##逻辑值的计数和比例：sum(x > 10)、mean(y ==10)


###按多个变量分组
#每次会用掉一个分组变量
daily <- group_by(flights, year, month, day)
(per_day <- summarize(daily, flights = n()))
(per_month <- summarize(per_day, flights = sum(flights)))
(per_year <- summarize(per_month, flights = sum(flights)))
##取消分组
daily %>%
  ungroup() %>%
  summarize(flights = n())


### 分组新变量和筛选器
#找出每个分组中最差的成员
flights_sml %>%
  group_by(year, month, day) %>%
  filter(rank(desc(arr_delay)) < 10)
#找出大于某个阈值的所有分组
(popular_dests <- flights %>%
  group_by(dest) %>%
  filter(n() > 365))