R-tidyverse系列-dplyr表格操作

stringr字符处理 - (jianshu.com)
dplyr表格操作 - (jianshu.com)

对于表格来说，一般行(row)代表不同的样本，列(column)代表不同角度的观测值。tibble是dplyr默认支持的一种表格类型，相比于R常规的data.frame表格类型有些许不同。

本次基于dplyr包，学习表格操作的知识点有：
1、表格数据统计
2、表格筛选
3、表格新增列
4、表格排序
5、表格的列名/行名
6、两个表格合并

library(tidyverse)
# -- Attaching packages ----------------------------------------------------- tidyverse 1.3.1 --
# √ ggplot2 3.3.5     √ purrr   0.3.4
# √ tibble  3.1.2     √ dplyr   1.0.7
# √ tidyr   1.1.3     √ stringr 1.4.0
# √ readr   2.0.0     √ forcats 0.5.1
# -- Conflicts -------------------------------------------------------- tidyverse_conflicts() --
# x dplyr::filter() masks stats::filter()
# x dplyr::lag()    masks stats::lag()

0、示例数据

mtcars %>% head
#                     mpg cyl disp  hp drat    wt  qsec vs am gear carb
# Mazda RX4         21.0   6  160 110 3.90 2.620 16.46  0  1    4    4
# Mazda RX4 Wag     21.0   6  160 110 3.90 2.875 17.02  0  1    4    4
# Datsun 710        22.8   4  108  93 3.85 2.320 18.61  1  1    4    1
# Hornet 4 Drive    21.4   6  258 110 3.08 3.215 19.44  1  0    3    1
# Hornet Sportabout 18.7   8  360 175 3.15 3.440 17.02  0  0    3    2
# Valiant           18.1   6  225 105 2.76 3.460 20.22  1  0    3    1

iris %>% head
# # A tibble: 6 x 5
# Sepal.Length Sepal.Width Petal.Length Petal.Width Species
#                          
# 1          5.1         3.5          1.4         0.2 setosa 
# 2          4.9         3            1.4         0.2 setosa 
# 3          4.7         3.2          1.3         0.2 setosa 
# 4          4.6         3.1          1.5         0.2 setosa 
# 5          5           3.6          1.4         0.2 setosa 
# 6          5.4         3.9          1.7         0.4 setosa

1、表格数据统计

• 表格数据的统计一般指对感兴趣的列进行例如均值、最值等统计指标的计算。
• dplyr支持在表征分组信息列的基础上，对其它列进行统计；搭配管道符%>%操作尤其强大
  关于管道操作符，可参考上一小节

1.1 group_by()

#得到具有Groups属性的tibble数据格式
by_cyl <- mtcars %>% group_by(cyl)
by_cyl  %>% head   #如下表示根据cyl分为3组，但注释数据框本身信息没有任何丢失
# # A tibble: 6 x 11
# # Groups:   cyl [3]
# mpg   cyl  disp    hp  drat    wt  qsec    vs    am  gear  carb
#           
# 1  21       6   160   110  3.9   2.62  16.5     0     1     4     4
# 2  21       6   160   110  3.9   2.88  17.0     0     1     4     4
# 3  22.8     4   108    93  3.85  2.32  18.6     1     1     4     1
# 4  21.4     6   258   110  3.08  3.22  19.4     1     0     3     1
# 5  18.7     8   360   175  3.15  3.44  17.0     0     0     3     2
# 6  18.1     6   225   105  2.76  3.46  20.2     1     0     3     1

mtcars %>% group_by(cyl, am) #根据两个分类列进行分组
# A tibble: 32 x 11
# Groups:   cyl, am [6]

1.2 summarise()

summarise(mtcars, mean_disp = mean(disp))
mtcars %>% summarise(mean_disp = mean(disp))
#   mean_disp
# 1  230.7219

• 如上是最简单的用法。可以对分组后的tibble进行高效的统计操作；并且可以通过?summarise查看有哪些常用的统计函数

mtcars %>%
  group_by(cyl) %>%  #分组统计，再纵向合并分组结果
  summarise(mean = mean(disp), n = n()) 
# # A tibble: 3 x 3
# cyl  mean     n
#   
# 1     4  105.    11
# 2     6  183.     7
# 3     8  353.    14

# n()函数表示每个分类水平的数目，类似于table()。一般建议在分组统计时都添加下，对数据有整体的把握


mtcars %>%
  group_by(cyl) %>%
  summarise(qs = quantile(disp, c(0.25, 0.75)), prob = c(0.25, 0.75), test = c("aa","bb"))
# # A tibble: 6 x 4
# # Groups:   cyl [3]
# cyl    qs  prob test 
#    
# 1     4  78.8  0.25 aa   
# 2     4 121.   0.75 bb   
# 3     6 160    0.25 aa   
# 4     6 196.   0.75 bb   
# 5     8 302.   0.25 aa   
# 6     8 390    0.75 bb’


# 搭配across同时对多列进行统计
iris %>%
  group_by(Species) %>% #自动变为tibble
  summarise(across(starts_with("Sepal"), list(mean = mean, sd = sd)))
iris %>%
  group_by(Species) %>% 
  summarise(across(starts_with("Sepal"), mean, .names = "mean_{.col}"))

1.3 n() 与 count()

#下面两个命令等价
mtcars %>% count(cyl)
mtcars %>% 
  group_by(cyl) %>%
  summarise(n = n())
# # A tibble: 3 x 2
# cyl     n
#  
# 1     4    11
# 2     6     7
# 3     8    14


#进阶用法
mtcars %>% count(cyl, sort = T) #降序排列
mtcars %>% count(cyl, vs)  #两个分组变量
mtcars %>% count(qsec_int=round(qsec))  #计算新的一列，并分组统计
mtcars %>% add_count(cyl) #在原有表格基础上增添一列，为相应level的counts总数

2、表格筛选

2.1 filter()筛选行

• 按列的属性值筛选感兴趣的样本

mtcars %>% filter(cyl == 6)
mtcars %>% filter(cyl == 6, disp > 200)  #与
mtcars %>% filter(cyl == 6 & disp > 200) #与
mtcars %>% filter(cyl == 6 | disp > 200) #或


# 搭配group_by()，分组筛选
# 这在基于全组水平的筛选指标很有意义
mtcars %>% 
  group_by(vs) %>%
  filter(disp == max(disp)) #筛选按 vs 分组里，disp的最大值
mtcars %>% 
  group_by(vs) %>%
  filter(disp > median(disp)) #筛选按 vs 分组里，disp 里的前50%

# top_n() 函数可更加方便的选择每组的top X、tail X
# top_n(x, n, wt)
mtcars %>% 
  group_by(vs) %>%
  top_n(3, wt = disp) #top X
mtcars %>% 
  group_by(vs) %>%
  top_n(-3, wt = disp) #tail X

# top_frac(x, n, wt)

2.2 distinct()与n_distinct 行去重

df <- tibble(
  x = sample(3, 10, rep = TRUE),
  y = sample(3, 10, rep = TRUE)
)
# # A tibble: 10 x 2
# x     y
#  
# 1     3     1
# 2     1     1
# 3     2     1
# 4     1     2
# 5     3     2
# 6     3     2
# 7     2     2
# 8     2     3
# 9     3     1
# 10     3     3

nrow(df)
df %>% n_distinct() #有多少完全unique的行
# df %>% distinct() %>% nrow()
nrow(distinct(df)) #默认去掉完全一致的行
nrow(distinct(df, x)) #按照指定的一/几列的值进行去重

2.3 挑选列

mtcars %>% select(mpg)
mtcars %>% select(mpg, am)
mtcars %>% select(mpg:hp)
mtcars %>% select(starts_with("c")) #列名以p开头的
mtcars %>% select(contains("p")) #列名含有 p 的
mtcars %>% select(starts_with("c") | contains("p") | vs) 
mtcars %>% select(!mpg)  #反选
mtcars %>% select(-mpg)  #反选
mtcars %>% select(hp, everything())  #把指定列放在第一列

3、表格新增列

3.1 mutate 在原有表格基础上增添新的列

#根据已有的列计算得到新的列
starwars %>%
  select(name, mass) %>%
  mutate(
    mass2 = mass * 2,
    mass2_squared = mass2 * mass2
  )

 #修改已有的列
starwars %>%
  select(name, height, mass, homeworld) %>%  
  mutate(
    mass = NULL,  #删除原有列
    height = height * 0.0328084 # 修改原有的列
  )

#多列批量操作
starwars %>%
  select(name, homeworld, species) %>%
  mutate(across(!name, as.factor))   
iris %>%
  as_tibble() %>%
  mutate(across(c(Sepal.Length, Sepal.Width), round))

#涉及到总体水平时，group_by分组会有影响
starwars %>%
  select(name, mass, species) %>%
  mutate(mass_norm = mass / mean(mass, na.rm = TRUE))
starwars %>%
  select(name, mass, species) %>%
  group_by(species) %>%   
  mutate(mass_norm = mass / mean(mass, na.rm = TRUE))

#新增列，表示组内排名
starwars %>%
  select(name, mass, homeworld) %>%
  group_by(homeworld) %>%
  mutate(rank = min_rank(mass))  #分组排名，默认为升序，即值越小，排名越前
starwars %>%
  select(name, mass, homeworld) %>%
  group_by(homeworld) %>%
  mutate(rank = min_rank(desc(mass)))  #分组排名，设置为降序，即值越大，排名越前



###关于各种rank###
x <- c(5, 1, 3, 2, 2)
rank(x)
#[1] 5.0 1.0 4.0 2.5 2.5
row_number(x)   # rank no duplication；可能不公平；相同情况下，排在前面的排名高
#[1] 5 1 4 2 3
min_rank(x)    # rank no duplication；有相同排名，会有gap的情况
dense_rank(x)   #rank no duplication；有相同排名，没有gap的情况
percent_rank(x) #a number between 0 and 1 computed by rescaling min_rank to [0, 1]
cume_dist(x) #a cumulative distribution function. Proportion of all values less than or equal to the current rank.
#################


# 默认新增的列会在原始表格的最右边
# Experimental: you can override with `.before` or `.after`
df <- tibble(x = 1, y = 2)
df %>% mutate(z = x + y)
df %>% mutate(z = x + y, .before = 3)  #指明新增的列在第几列
df %>% mutate(z = x + y, .after = x)   #指明新增的列放在哪一列的后面

?transmute   #只输出新增的列
df %>% transmute(z = x + y)
df %>% mutate(z = x + y, .keep = "none")  # .keep = 'all'    the default

4、表格排序

• arrange()

mtcars %>% head
mtcars %>% arrange(disp)
mtcars %>% arrange(desc(disp))  # desc 降序排列
arrange(mtcars, carb, disp)  #先按carb升序排列，再按disp升序排列

mtcars %>% count(cyl, sort = T) #降序排列
mtcars %>% 
  count(cyl) %>%
  arrange(n)  #升序排列

5、表格的列名/行名

• rename: 修改列名

#修改列名
iris <- as_tibble(iris) # so it prints a little nicer
rename(iris, petal_length = Petal.Length)   #  new = old

• rownames_to_column、column_to_rownames行名与列的转换

mtcars %>% head
#行名变为列
rownames_to_column(mtcars, var = "car") %>% head
mtcars %>% 
  rownames_to_column(var = "car") %>% head
mtcars %>% 
  rownames_to_column(var = "car") %>%
#指定列变为行名
  column_to_rownames(var = "car") %>% head

6、两个表格合并

• 示例表格数据

df1 <- data.frame(x1 = c(1, 2, 3, 4, 5, 6),
                    y1 = c("a", "b", "c", "d","e","f"))
#   x1 y1
# 1  1  a
# 2  2  b
# 3  3  c
# 4  4  d
# 5  5  e
# 6  6  f

df2 <- data.frame(x2 = c(1, 2, 3, 4, 5, 6),
                    y2 = c("q", "w", "e", "r","t","y"))
#   x2 y2
# 1  1  q
# 2  2  w
# 3  3  e
# 4  4  r
# 5  5  t
# 6  6  y

df3 <- data.frame(x3 = c( 2, 4, 5, 6),
                    y3 = c( "a", "s", "d","f"))
#   x3 y3
# 1  2  a
# 2  4  s
# 3  5  d
# 4  6  f

根据不同的分析目的，有多种合并方法~

6.1、 内连接

• 取交集，即保留同时在两个表中的观测

# merge(df1, df3, by.x = "y1", by.y = "y3")
inner_join(df1, df3, by=c("y1"="y3"))
#   x1 y1 x3
# 1  1  a  2
# 2  4  d  5
# 3  6  f  6

注意：如果这两个列的名相同，合并时直接交代行名即可，例如inner_join(df1_1, df3_1, by="y")，merge(df1_1, df3_1, by = "y"),下同~

6.2、外连接

• （1）左连接--保留左边表格的所有观测，缺失值用NA值代替

# merge(df1, df3, by.x = "y1", by.y = "y3", all.x = T)
left_join(df1, df3, by=c("y1"="y3"))
#   x1 y1 x3
# 1  1  a  2
# 2  2  b NA
# 3  3  c NA
# 4  4  d  5
# 5  5  e NA
# 6  6  f  6

• （2）右连接--保留右边表格的所有观测，缺失值用NA值代替

# merge(df1, df2, by.x = "y1", by.y = "y2", all.y = T)
right_join(df1, df2, by=c("y1"="y2"))
#   x1 y1 x2
# 1  5  e  3
# 2 NA  q  1
# 3 NA  w  2
# 4 NA  r  4
# 5 NA  t  5
# 6 NA  y  6

• （3）外连接--保留两个表格里的所有观测，缺失值用NA值代替

# merge(df1, df2, by.x = "y1", by.y = "y2", all.x = T, all.y = T)
full_join(df1, df2, by=c("y1"="y2"))
#    x1 y1 x2
# 1   1  a NA
# 2   2  b NA
# 3   3  c NA
# 4   4  d NA
# 5   5  e  3
# 6   6  f NA
# 7  NA  q  1
# 8  NA  w  2
# 9  NA  r  4
# 10 NA  t  5
# 11 NA  y  6