R在频率表中保留零长度组

Question

我有以下数据集:

gender<-c('male' ,'male', 'male', 'male','male',
          'female', 'female', 'female','female', 'female' ,'female', 'female', 'female','female')   
clothes<-c('hat', 'hat', 'hat', 'shirt', 'shirt', 'hat', 'hat', 'hat', 'shirt', 'shirt', 'shirt', 'dress', 'dress', 'dress')
color<-c('blue', 'blue', 'green', 'blue', 'brown', 'green', 'brown', 'brown', 'blue', 'green', 'green', 'blue', 'green', 'green')
x<-data.frame(gender, clothes, color)

我需要按颜色按衣服制作性别频率表，NA 仅用于缺少颜色。性别和所有衣服级别都应与 3 个颜色级别相关联。但是，对于一个性别级别，我缺少衣服级别的“衣服”，并且我不希望它填充 NA，我希望将其完全省略。

我试着计算:

x$color<-as.factor(x$color)
x_agg<-x%>%
  group_by(gender, clothes, color)%>%
  tally()

而这根本达不到目的；对于缺少任何变量的级别，我没有得到 NA。

当我应用以下代码时:

library(tidyverse)
x_agg<-x%>%
  group_by(gender, clothes, color)%>%
  summarise(cnt=n())%>%
  ungroup() %>%
  complete(gender, clothes, color,
           fill = list(N = 0))

我得到男性的 NA - 连衣裙 - 所有颜色。但我想要的只是最后一个分组变量(颜色)的 NA，而不是衣服和颜色。像这样:

gender<-c('male' ,'male', 'male', 'male','male','male',
      'female', 'female', 'female','female', 'female' ,'female', 'female', 'female','female')   
clothes<-c('hat', 'hat', 'hat', 'shirt', 'shirt', 'shirt', 
       'hat', 'hat', 'hat', 'shirt', 'shirt', 'shirt', 'dress', 'dress', 'dress')
color<-c('blue', 'green', 'brown', 
     'blue', 'green', 'brown',
     'blue', 'green', 'brown',
     'blue', 'green', 'brown',
     'blue', 'green', 'brown')
cnt<-c(2, 1, NA, 1, NA, 1, NA, 1, 2, 1, 2, NA, 1, 2, NA)     
x_agg1<-data.frame(gender, clothes, color, cnt)     

或者这是一张图片:


我想我尝试了我能想到的一切。有关于堆栈溢出的建议，但都与仅按一个变量分组或为每个单个分组变量的所有级别填充 NA 相关。但是如果只需要填充/保留一个变量的级别，则不清楚该怎么办？有什么建议？

Answer 1

而不是 group_by/summarise ，我们也可以做count

library(dplyr)
library(tidyr)
x %>%
     count(gender, clothes, color) %>%
     group_by(gender, clothes) %>%
     complete(color)
# A tibble: 15 x 4
# Groups:   gender, clothes [6]
#   gender clothes color     n
#   <fct>  <fct>   <fct> <int>
# 1 female dress   blue      1
# 2 female dress   brown    NA
# 3 female dress   green     2
# 4 female hat     blue     NA
# 5 female hat     brown     2
# 6 female hat     green     1
# 7 female shirt   blue      1
# 8 female shirt   brown    NA
# 9 female shirt   green     2
#10 male   hat     blue      2
#11 male   hat     brown    NA
#12 male   hat     green     1
#13 male   shirt   blue      1
#14 male   shirt   brown     1
#15 male   shirt   green    NA