节省tidyverse中的管道%>%

分享{dplyr}中一些鲜为人知的功能，喜欢可以关注我的公众号R语言数据分析指南

library(tidyverse)
data("penguins", package = "palmerpenguins")
penguins <- na.omit(penguins)

1.rename()里面select()

penguins %>% 
  select(species, island) %>% 
  rename(penguin_species = species)

penguins %>% 
  select(penguin_species = species,
         island)

2.rename()里面count()

penguins %>% count(species) %>% 
  rename(total = n)

penguins %>% 
  count(species, name = "total")

还可以使用与select()上面的示例来计数并重新命名

penguins %>% 
  count(species) %>% 
  rename(total = n,penguin_species = species)

penguins %>% 
  count(penguin_species = species, name = "total")

请注意，传递给name参数的新名称必须用引号引起来，但选定列的新名称不必用引号引起来

3.mutate()里面count()

这很简单-您只需在mutate()内部执行以下操作即可count()

penguins %>% 
  mutate(long_beak = bill_length_mm > 50) %>% 
  count(long_beak)

penguins %>% 
  count(long_beak = bill_length_mm > 50)

当然，当指定多个变量以下列方式计数时，此方法也适用

penguins %>% 
  mutate(long_beak = bill_length_mm > 50,
         is_adelie = species == "Adelie") %>% 
  count(is_adelie, long_beak)

penguins %>% 
  count(long_beak = bill_length_mm > 50,
        is_adelie = species == "Adelie")

4. transmute()+select()

penguins %>% 
mutate(body_mass_kg = body_mass_g/1000) %>% 
select(body_mass_kg)

penguins %>% 
transmute(body_mass_kg = body_mass_g/1000)

transmute()过去我很少使用过，因为我认为它只能返回经过修改的列，这将是非常有限的（例如，在上面的示例中，以公斤为单位的企鹅体重有什么好处？）

但是实际上，您只可以命名要包括的列，transmute()就像select()继承未修改的列一样。当然，您可以在执行操作时对其“重命名

penguins %>% 
  mutate(body_mass_kg = body_mass_g/1000) %>% 
  select(species, island, body_mass_kg) %>% 
  rename(penguin_species = species)

penguins %>% 
  transmute(penguin_species = species,
            island,
            body_mass_kg = body_mass_g/1000)

5.ungroup()里面summarize()

penguins %>% 
  group_by(island, species) %>% 
  summarize(mean_mass = mean(body_mass_g, na.rm = TRUE)) %>% 
  ungroup()

因为summarize()仅按defaut删除最后一个分组变量，这意味着如果ungroup()不调用island，输出仍按变量分组：

penguins %>% 
  group_by(island, species) %>% 
  summarize(mean_mass = mean(body_mass_g, na.rm = TRUE)) %>% 
  group_vars()

penguins %>% 
  group_by(island, species) %>% 
  summarize(mean_mass = mean(body_mass_g, na.rm = TRUE)) %>% 
  ungroup() %>% 
  group_vars()

也可以简单地设置.groups参数内summarize()，为'drop'达到相同的：

penguins %>% 
  group_by(island, species) %>% 
  summarize(mean_mass = mean(body_mass_g, na.rm = TRUE), .groups = 'drop')

  # A tibble: 5 x 3
    island    species   mean_mass
    <fct>     <fct>         <dbl>
  1 Biscoe    Adelie        3710.
  2 Biscoe    Gentoo        5092.
  3 Dream     Adelie        3701.
  4 Dream     Chinstrap     3733.
  5 Torgersen Adelie        3709.

6. arrange()+其他功能slice()

如果您想获取按列排序的前n行，则可以使用top_n()，它提供了一种更简单的方式 slice()+arrange()：

penguins %>% 
  arrange(desc(body_mass_g)) %>% 
  slice(1:5)

penguins %>% 
  top_n(5, wt = body_mass_g)

penguins %>% 
  top_n(5, wt = body_mass_g)

penguins %>% 
  slice_max(order_by = body_mass_g, n = 5)

新slice_*()功能最重大的变化是为分组数据帧添加了适当的行为

例如，下面的示例返回每种物种的重量百分比最高的5％的企鹅：

penguins %>% 
  group_by(species) %>% 
  slice_max(body_mass_g, prop = .05)

7.按组进行计数和求和 add_count()

add_count() 添加一列，其中包含每个组（或组的组合）的计数

##### Long Form #####
# penguins %>% 
#   group_by(species) %>% 
#   mutate(count_by_species = n()) %>% 
#   ungroup()

penguins %>% 
  add_count(species, name = "count_by_species") %>% 
  select(-contains("mm"))

 # A tibble: 333 x 6
     species island    body_mass_g sex     year count_by_species
     <fct>   <fct>           <int> <fct>  <int>            <int>
   1 Adelie  Torgersen        3750 male    2007              146
   2 Adelie  Torgersen        3800 female  2007              146
   3 Adelie  Torgersen        3250 female  2007              146
   4 Adelie  Torgersen        3450 female  2007              146
   5 Adelie  Torgersen        3650 male    2007              146
   6 Adelie  Torgersen        3625 female  2007              146
   7 Adelie  Torgersen        4675 male    2007              146
   8 Adelie  Torgersen        3200 female  2007              146
   9 Adelie  Torgersen        3800 male    2007              146
  10 Adelie  Torgersen        4400 male    2007              146
  # ... with 323 more rows

可以使用wt来按组有效地获取总和（也许有点笨拙，但非常有用)

##### Long Form #####
# penguins %>% 
#   group_by(species) %>% 
#   mutate(total_weight_by_species = sum(body_mass_g)) %>% 
#   ungroup()
  
penguins %>% 
  add_count(species, wt = body_mass_g,
  name ="total_weight_by_species") %>% 
  select(-contains("mm"))

 # A tibble: 333 x 6
     species island    body_mass_g sex     year total_weight_by_species
     <fct>   <fct>           <int> <fct>  <int>                   <int>
   1 Adelie  Torgersen        3750 male    2007                  541100
   2 Adelie  Torgersen        3800 female  2007                  541100
   3 Adelie  Torgersen        3250 female  2007                  541100
   4 Adelie  Torgersen        3450 female  2007                  541100
   5 Adelie  Torgersen        3650 male    2007                  541100
   6 Adelie  Torgersen        3625 female  2007                  541100
   7 Adelie  Torgersen        4675 male    2007                  541100
   8 Adelie  Torgersen        3200 female  2007                  541100
   9 Adelie  Torgersen        3800 male    2007                  541100
  10 Adelie  Torgersen        4400 male    2007                  541100
  # ... with 323 more rows

默认情况下，add_tally()添加行数，您已经可以使用mutate(n = n())进行处理

penguins %>% 
  add_count(species, wt = body_mass_g, 
name = "total_weight_by_species") %>% 
  add_tally(wt = body_mass_g,
name = "total_weight_of_all_species") %>% 
  select(1:2, last_col(0):last_col(1))

  # A tibble: 333 x 4
     species island    total_weight_of_all_species total_weight_by_species
     <fct>   <fct>                           <int>                   <int>
   1 Adelie  Torgersen                     1400950                  541100
   2 Adelie  Torgersen                     1400950                  541100
   3 Adelie  Torgersen                     1400950                  541100
   4 Adelie  Torgersen                     1400950                  541100
   5 Adelie  Torgersen                     1400950                  541100
   6 Adelie  Torgersen                     1400950                  541100
   7 Adelie  Torgersen                     1400950                  541100
   8 Adelie  Torgersen                     1400950                  541100
   9 Adelie  Torgersen                     1400950                  541100
  10 Adelie  Torgersen                     1400950                  541100
  # ... with 323 more rows